[House Hearing, 109 Congress]
[From the U.S. Government Publishing Office]
SCADA SYSTEMS AND THE TERRORIST
THREAT: PROTECTING THE NATION'S
CRITICAL CONTROL SYSTEMS
=======================================================================
JOINT HEARING
before the
SUBCOMMITTEE ON ECONOMIC
SECURITY, INFRASTRUCTURE
0PROTECTION, AND CYBERSECURITY
with the
SUBCOMMITTEE ON EMERGENCY
PREPAREDNESS, SCIENCE, AND TECHNOLOGY
of the
COMMITTEE ON HOMELAND SECURITY
HOUSE OF REPRESENTATIVES
ONE HUNDRED NINTH CONGRESS
FIRST SESSION
__________
OCTOBER 18, 2005
__________
Serial No. 109-45
__________
Printed for the use of the Committee on Homeland Security
[GRAPHIC] [TIFF OMITTED] TONGRESS.#13
Available via the World Wide Web: http://www.gpoaccess.gov/congress/
index.html
__________
U.S. GOVERNMENT PRINTING OFFICE
31-506 WASHINGTON : 2007
_____________________________________________________________________________
For Sale by the Superintendent of Documents, U.S. Government Printing Office
Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; (202) 512�091800
Fax: (202) 512�092250 Mail: Stop SSOP, Washington, DC 20402�0900012007
COMMITTEE ON HOMELAND SECURITY
Peter T. King, New York, Chairman
Don Young, Alaska Bennie G. Thompson, Mississippi
Lamar S. Smith, Texas Loretta Sanchez, California
Curt Weldon, Pennsylvania Edward J. Markey, Massachusetts
Christopher Shays, Connecticut Norman D. Dicks, Washington
John Linder, Georgia Jane Harman, California
Mark E. Souder, Indiana Peter A. DeFazio, Oregon
Tom Davis, Virginia Nita M. Lowey, New York
Daniel E. Lungren, California Eleanor Holmes Norton, District of
Jim Gibbons, Nevada Columbia
Rob Simmons, Connecticut Zoe Lofgren, California
Mike Rogers, Alabama Sheila Jackson-Lee, Texas
Stevan Pearce, New Mexico Bill Pascrell, Jr., New Jersey
Katherine Harris, Florida Donna M. Christensen, U.S. Virgin
Bobby Jindal, Louisiana Islands
Dave G. Reichert, Washington Bob Etheridge, North Carolina
Michael McCaul, Texas James R. Langevin, Rhode Island
Charlie Dent, Pennsylvania Kendrick B. Meek, Florida
Ginny Brown-Waite, Florida
______
Subcommittee on Economic Security, Infrastructure Protection, and
Cybersecurity
Daniel E. Lungren, California, Chairman
Don Young, Alaska Loretta Sanchez, California
Lamar S. Smith, Texas Edward J. Markey, Massachusetts
John Linder, Georgia Norman D. Dicks, Washington
Mark E. Souder, Indiana Peter A. DeFazio, Oregon
Mike Rogers, Alabama Zoe Lofgren, California
Stevan Pearce, New Mexico Sheila Jackson-Lee, Texas
Katherine Harris, Florida Bill Pascrell, Jr., New Jersey
Bobby Jindal, Louisiana James R. Langevin, Rhode Island
Peter T. King, New York (Ex Bennie G. Thompson, Mississippi
Officio) (Ex Officio)
______
SUBCOMMITTE ON EMERGENCY PREPAREDNESS, SCIENCE, AND TECHNOLOGY
Dave G. Reichert, Washington, Chairman
Lamar S. Smith, Texas Bill Pascrell, Jr., New Jersey
Curt Weldon, Pennsylvania Loretta Sanchez, California
Rob Simmons, Connecticut Norman D. Dicks, Washington
Mike Rogers, Alabama Jane Harman, California
Stevan Pearce, New Mexico Nita M. Lowey, New York
Katherine Harris, Florida Eleanor Holmes Norton, District of
Michael McCaul, Texas Columbia
Charlie Dent, Pennsylvania Donna M. Christensen, U.S. Virgin
Ginny Brown-Waite, Florida Islands
Peter T. King, New York (Ex Bob Etheridge, North Carolina
Officio) Bennie G. Thompson, Mississippi
(Ex Officio)
(II)
C O N T E N T S
----------
Page
STATEMENTS
The Honorable Daniel E. Lungren, a Representative in Congress
From the State of California, and Chairman, Subcommittee on
Economic Security, Infrastructure Protection, and
Cybersecurity:
Oral Statement................................................. 1
Prepared Statement............................................. 1
The Honorable Loretta Sanchez, a Representative in Congress From
the State of California, and Ranking Member, Subcommittee on
Economic Security, Infrastructure Protection, and Cybersecurity 2
The Honorable Dave G. Reichert, a Representative in Congress From
the State of Washington, and Chairman, Subcommittee on
Emergency Preparedness, Science, and Technology:
Oral Statement................................................. 2
Prepared Statement............................................. 3
The Honorable Bill Pascrell, Jr., a Representative in Congress
From the State of New Jersey, and Ranking Member, Subcommittee
on Emergency Preparedness, Science and Technology:
Prepared Statement............................................. 3
The Honoralee Peter T. King, a Representative in Congress From
the State of New York, and Chairman, Committee on Homeland
Security:
Prepared Statement............................................. 4
The Honorable Bennie G. Thompson, a Representative in Congress
From the State of Mississippi, and Ranking Member, Committee on
Homeland Security:
Oral Statement................................................. 5
Prepared Statement............................................. 58
The Honorable Donna M. Christensen, a Delegate in Congress From
the U.S. Virgin Islands........................................ 67
The Honorable Norman D. Dicks, a Representative in Congress From
the State Washington........................................... 68
The Honorable Bob Etheridge, a Representative in Congress From
the State of North Carolina.................................... 65
The Honorable Sheila Jackson-Lee, a Representative in Congress
From the State of Texas........................................ 64
The Honorable Eleanor Holmes Norton, a Delegate in Congress From
the District of Columbia....................................... 62
The Honorable Stevan Pearce, a Representative in Congress From
the State of New Mexico........................................ 56
The Honorable Ginny Brown-Waite, a Representative in Congress
From the State of Florida...................................... 60
Witnesses
Dr. K.P. Ananth, Associate Laboratory Director--National and
Homeland Security, Idaho National Laboratory:
Oral Statement................................................. 24
Prepared Statement............................................. 25
Mr. Alan Paller, Director of Research, The SANS Institute:
Oral Statement................................................. 40
Prepared Statement............................................. 42
Mr. Donald ``Andy'' Purdy, Acting Director, National Cyber
Security Division, U.S. Department of Homeland Security:
Oral Statement................................................. 6
Prepared Statement............................................. 7
Dr. William Rush, Institute Physicist, Gas Technology Institute:
Oral Statement................................................. 31
Prepared Statement............................................. 33
Mr. Larry Todd, Director, Security, Safety and Law Enforcement
Bureau of Reclamation, U.S. Department of the Interior:
Oral Statement................................................. 14
Prepared Statement............................................. 15
Dr. Sam Varnado, Director of Information Operations Center,
Sandia National Laboratory:
Oral Statement................................................. 16
Prepared Statement............................................. 18
APPENDIX
Dr. K.P. Ananth Responses to the Honorable Daniel E. Lungren
Questions...................................................... 71
Mr. Donald ``Andy'' Purdy Responses to the Honorable Bennie G.
Thompson Questions............................................. 81
Mr. Larry Todd Responses to the Honorable Bennie G. Thompson
Questions...................................................... 86
Dr. Sam Varnado Responses to the Honorable Bennie G. Thompson
Questions...................................................... 89
SCADA SYSTEMS AND THE TERRORIST
THREAT: PROTECTING THE NATION'S
CRITICAL CONTROL SYSTEMS
----------
Tuesday, October 18, 2005
U.S. House of Representatives,
Committee on Homeland Security,
Subcommittee on Economic Security,
Infrastructure Protection, and Cybersecurity,
with the
Subcommittee on Emergency
Preparedness, Science, and Technology,
Washington, DC.
The subcommittee met, pursuant to call, at 4 p.m., in Room
311, Cannon House Office Building, Hon. Dan Lungren [chairman
of the Subcommittee on Economic Security, Infrastructure
Protection, and Cybersecurity] presiding.
Present: Representatives Lungren, Reichert, Pearce, Brown-
Waite, Pascrell, Thompson, Dicks, Norton, Jackson-Lee,
Christensen, Etheridge and Sanchez.
Mr. Lungren. The joint hearing of the Committee on Homeland
Security Subcommittee on Economic Security, Infrastructure
Protection and Cybersecurity and the Subcommittee on Emergency
Preparedness, Science and Technology will come to order. The
subcommittees are meeting today in joint session to hear
testimony on supervisory control and data acquisition systems,
better known as SCADA systems, in the effort to protect these
critical control systems from terrorist attack.
We have been informed that we will have votes starting at
approximately 4:30, and as a result, we are going to have a
major interruption. We have six major witnesses here on a very
important matter, so I am going to not give my opening
statement. It will be included as a part of the record. And
then we will proceed.
Prepared Opening Statement of the Honorable Daniel Lungren
Good morning and I would like to welcome everyone to this joint
hearing of the Committee on Homeland Security's Subcommittee on
Economic Security, Infrastructure Protection, and Cybersecurity and the
Subcommittee on Emergency Preparedness, Science & Technology. I thank
Chairman Reichert and Ranking Member Pascrell for agreeing to hold this
jointly, as this critical issue has far reaching impacts.
We convene today to focus on the protection of control systems at
our Nation's critical infrastructure. Control systems are utilized in a
wide variety of industries--such as electrical generation and
distribution, oil and gas systems, traffic signals and other
transportation supervision, water management (including dams), and
manufacturing industries. These control systems are commonly referred
to as SCADA systems.
These computer terminals have the ability to give supervisory
control to a central user over separate and often disparate functions
or processes. Further, SCADA systems collect information from remote
locations and coalesce it into one location.
Now what does this actually mean? Simply put, a manufacturing
facility or any of the forementioned facilities incorporate many
different processes and functions. To safely, securely, and efficiently
run the facility, companies must be able to monitor and adjust these
processes simultaneously. Before SCADA systems, workers would be placed
throughout a facility and manually monitor and adjust the various
systems. SCADA systems bring monitoring and control of these functions
into one centralized location, making it easier and more efficient to
run these processes.
At the same time, these systems present serious security
challenges. Because these terminals control crucial systems within our
critical infrastructure and are often connected to networks and can be
remotely accessed, they present an attractive means for those wishing
to cause harm and confusion.
Securing SCADA systems is similar to securing all of our cyber
infrastructure; however, the consequences are potentially very
different. Minimally, adversaries could target SCADA systems through
cyber networks, utilizing common cyber attack methods to render the
SCADA systems unusable. This could slow down, stop, or endanger the
functions of the facility. This would result in not only serious
problems at that facility but potential cascading effects on other
facilities or processes that are dependent on the attacked facility.
Even worse, terrorists could utilize SCADA systems for their own
sinister motives--causing a pipeline to burst, opening flood gates on
dams, or shutting down our electric supply, all without ever gaining
access to the facility.
Part of this hearing will be to understand the function of these
systems within the greater picture of our critical infrastructure and
to understand the general vulnerabilities, consequences, and
interdependencies of these systems. Although there are literally
thousands of SCADA systems across the U.S., not all of these control
systems involve industries or facilities that would be considered high
risk.
The threat to these systems has long been recognized and the
Federal government, the private sector, and this country's best minds
have been working for years to address it. The second part of this
hearing then, is to understand what progress has been made--at all
levels--to address these vulnerabilities.
We have a diverse panel of experts today, representing the Federal
government, the National Labs, the dam industry, the gas industry, and
the cyber industry. I look forward to hearing from all of you about
your ongoing efforts, and your views on what we need to do to further
assist you in addressing SCADA security.
I am especially interested in hearing about the status of securing
our dams. We have seen recently in New Orleans what can happen when
nature overwhelms us, even with days of advance notice. The potential
consequences of an unanticipated attack could be far worse.
Again, I thank all of our witnesses for being here. I now recognize
the Ranking Member of the Subcommittee, the Gentle Lady from California
Ms. Sanchez, for any opening statement she'd like to make.
Mr. Lungren. The Chair would recognize the Ranking Minority
Member of the Subcommittee on Economic Security, Infrastructure
Protection and Cybersecurity, the gentlelady from California
Ms. Sanchez, for any statement she may make.
Ms. Sanchez. Thank you, Mr. Chairman. And considering I am
under the weather today and we are pushed against votes, I,
too, will hold my opening statement and submit it for the
record so that we can hear from the witnesses today. Thank you.
Mr. Lungren. I thank the gentlelady, and her prepared
statement will be made a part of the record.
The Chair would now recognize the Chairman of the
Subcommittee on Emergency Preparedness, Science and Technology,
the gentleman from Washington Mr. Reichert, for any statement
he may make.
Mr. Reichert. Thank you, Mr. Chairman. I, too, will
withhold boring you to death with my opening statement, and we
will ask
that it be placed in the record. Thank you, and welcome to the
witnesses today.
[The information follows:]
Prepared Opening Statement of the Honorable David Reichert
Thank you, Chairman Lungren. I would also like to welcome everyone,
especially our witnesses, to this joint hearing.
We are here today to discuss a topic that affects our everyday
lives, although many of us are never aware of it. Process and control
systems and the operations that they manage are critical to our Nation.
They enable us to have everything from clean drinking water and fuel
for our cars to electricity in our homes.
As a former law enforcement officer, I know firsthand that
prevention is the best way to save lives and protect property. So, I am
particularly interested in our Nation/s efforts to secure these
systems.
But, I also recognize that we can not expect to prevent every
attack, especially in an environment as open and free-flowing as
cyberspace. And, as we have seen in the aftermath of Hurricane Katrina,
our ability to recover from an incident--whether natural or manmade--
can be just as important as our ability to detect and prevent it from
happening in the first place.
Part of the mission of the Department of Homeland Security's
National Cyber Security Division is to ``establish a National
Cyberspace Response System.'' Ideally, such a system will rapidly
identify and respond to cyber incidents and help mitigate against any
damage caused by malicious cyberspace activities.
So far, we have fortunately not yet experienced a serious cyber
attack directed at the control systems that manage our Nation's
electrical grid, dams, and other critical plants. Undoubtedly, at some
point, our luck will run out. That is precisely why we must continue to
emphasize prevention and response and develop more robust SCADA
software technology.
I am, therefore, keenly interested in learning more about the
vulnerabilities of our SCADA systems, what the NCSD--in partnership
with the National labs and the private sector--has done to address such
vulnerabilities, and the additional steps that need to be taken to
establish and implement a cyber response system.
Again, I want to thank all our witnesses for being with us today. I
look forward to your testimony on this important issue.
Thank you, Mr. Chairman, and I yield back the balance of my time.
Mr. Lungren. All members of the committee--the Chairman
would recognize the Ranking Minority Member of the Subcommittee
on Emergency Preparedness, Science and Technology, the
gentleman from New Jersey Mr. Pascrell, for any statement he
might make. I would just inform the gentleman that we have all
waived our statements, but the gentleman may proceed as he
wishes.
Mr. Pascrell. I will waive it.
Mr. Lungren. Your statement will be made--a prepared
statement will be made a part of the record.
[The information follows:]
Prepared Statement of the Honorable Bill Pascrell, Jr.
I want to thank Chairman Lungren and Chairman Reichert for holding
a hearing on an issue of vital importance to our national security.
Indeed, protecting America's critical control systems is a topic
that, I believe, has not received the attention it deserves. We know
that vulnerabilities within these systems are abundant, and we know
that the threat of a terrorist attack against these systems is real.
Congress needs to engage in robust analysis and oversight in this
realm; we need to help ensure the security of the various control
systems that are used in critical infrastructure--and I am heartened
that today two Homeland Security subcommittees are leading the charge.
Obviously this is something that affects all of us. But as a
resident of New Jersey, I must say that this issue particularly
resonates with me.
There are a number of areas in my state, for example, that contain
key assets on which the region's economy and community functioning
depend--including critical utilities that provide gas, electric power,
water and telecommunications services.
A cyber attack on one of New Jersey's four nuclear power plants, or
100 chemical sites, for example, has the potential to be absolutely
devastating. Not only in terms of lives lost, but also in the regional
and national economic destruction it could bring forth. This is
serious, serious business.
Back in 2002, the National Infrastructure Protection Center
reported that a computer belonging to an individual who had links to
Osama bin Laden contained programs that clearly showed the individual's
interest in the structural engineering of various critical
infrastructures.
It also indicated that al-Qa'ida members had sought information
about control systems from multiple websites.
With this knowledge, one would assume that Washington would take
every appropriate step, take every possible measure, and institute
every conceivable action to ensure that critical infrastructure would
be greater protected.
Inexplicably, this doesn't seem to be the case.
In fact, DHS as a whole has been slow in completing its critical
infrastructure protection policies.
In December 2003, President Bush issued Presidential Directive 7,
establishing a national policy for federal departments and agencies to
prioritize critical infrastructure. DHS was charged with developing the
National Infrastructure Protection Plan (N.I.P.P.) to serve as the
guide for protecting infrastructure.
The N.I.P.P. was due in December 2004. In February 2005, an
``Interim plan'' was issued, setting a deadline of November 2005 for
the final plan. According to the GAO, the interim plan was incomplete:
it lacked both national-level milestones and sector-specific security
plans.
The plan remains incomplete to this day. We can't even get
proposals ready in a timely matter. This is unconscionable.
I'm also seriously concerned that the Department is not devoting
enough manpower to this threat. According to an August 12th response by
DHS to a request made by committee staff, there was only one full time
employee staffed exclusively to control system projects at the National
Cyber Security Division in the department.
One person. Surely it takes more than a single, lonely individual
to effectively coordinate the public and private efforts in the control
systems field?
The fact is this: the threats and dangers to control systems are
increasing.
Standardized technologies currently being used have commonly known
vulnerabilities allowing for easy exploitation. The connectivity of
control systems to other networks offers additional beaches in
security. Widespread public availability of technical information about
control systems continues to present a serious risk.
And the federal government isn't ready.
I look forward to the testimony from our witnesses today, and I
hope that this hearing is the first in a series of actions our
committee takes to ensure that control systems are as safe as they
possibly can be.
Mr. Lungren. All Members are reminded that opening
statements may be submitted for the record.
[The information follows:]
Prepared Opening Statement of the Honorable Peter King
Thank you. And thanks to our witnesses for appearing before these
Subcommittees today.
As Chairman Lungren pointed out-SCADA systems are an integral part
of our critical infrastructure. These real time control systems operate
our major industries that we rely on everyday, including our gas,
water, electric and oil facilities. They are integral parts of our
efficient operation of these industries- and our National economy.
SCADA systems control integral and vital processes of our
infrastructure with potential significant physical and public health
and ramifications if they are shut down or misused. SCADA systems are
part of the larger issue of cybersecurity and a vital component of
critical infrastructure protection.
Because these systems are connected to the internet or our
telephone network--these systems can be remotely accessed, and they are
easily penetrated. These systems were created decades ago and were
designed before security was as great a concern as it is now. Many
systems are not protected by basic security features, such as passwords
or firewalls.
The good news is that there have been no reported terrorist cyber-
attacks on domestic critical infrastructure control systems that have
resulted in significant damage. This does not indicate that it is not
possible or that terrorists are not interested in these
vulnerabilities. There are reports that al-Qa'ida computers found in
Afghanistan contained information on structural analysis programs for
dams and that these computers were used to search for information on
SCADA systems specifically.
There have been cases of non-terrorist individuals breaking into
control systems and in some cases causing damage including an instance
in Australia, in 2000 where a malicious former employee remotely
accessed the control system of a sewage plant and discharged almost
265,000 gallons of sewage into the local environment.
There are two things that we need to see happen. We need to be
working with industry and the National Labs to develop new secure
systems that can be put in as replacements or for new industries. But
we can not expect all of the owners and operators of SCADA systems to
incur the expensive cost of replacing existing control systems. Rather,
the second thing we need to see- is procedures and protocols developed
and distributed that can improve the security of these critical
systems. Utilizing encryption, installing security software on outdated
systems, training and educating employees on basic security procedures,
these things can be done to reduce the vulnerabilities without entirely
replacing the systems themselves.
I look forward to hearing from this panel on their thoughts on
these issues and what they have done specifically to improve the
security of the existing SCADA systems and the new SCADA systems being
produced. I know that DHS has worked with the National Labs and the
Dept. of Energy to develop programs to test existing systems, to model
interdependencies and vulnerabilities--but it is also evident that the
private sector has not waited for the Federal government to provide
guidance. I look forward to hearing from our private sector witnesses
as well, as to their efforts to secure this vital component of our
National infrastructure.
Thank you again, and I look forward to your testimony and the
opportunity to ask you questions.
Prepared Opening Statement of the Honorable Bennie G. Thompson
Thank you Mr. Chairman, Ranking Member Sanchez. I am glad we are
here today to consider this important issue.
SCADA systems perform vital functions in running much of our
industrial and critical infrastructure processes.
As technology continues to develop, this country will become more
reliant on computerized control systems to perform these vital
monitoring functions.
It is imperative that the Congress and this Administration act
quickly to solve the serious security problems that plague SCADA and
control systems.
The possibilities of a terrorist breaching a SCADA system are
incredibly frightening.
Nuclear power plants--like the one located in Port Gibson,
Mississippi, in my District--can potentially be at risk.
Electric grids, water management systems, and oil and gas control
systems are also all at risk. Attacks can result in unquantifiable
losses of infrastructure, money, and lives.
The risks to control systems posed by a natural disaster, like
Hurricane Katrina, must also be considered.
The hurricane shut down the electrical grid along the Gulf Coast,
thereby forcing two critical pipelines to shut down.
We're all still paying at the gas pump partially because of that
failure.
we spent the time, money, and energy building our critical
infrastructure systems; we must now spend the time, money, and energy
to protect them.
As you all know, protecting SCADA and control systems requires a
commitment from two entities.
The private sector must continue to identify current security
risks, modify and adopt new encryption standards, and create new
technologies to secure future systems.
It's also important for us here in Congress to determine what role
the federal government should play.
Should we provide incentives for SCADA systems to comply with best
practices? Should we establish new guidelines for existing SCADA
systems?
Should we use the leverage the federal government has when buying
SCADA systems for itself in order to create changes across the market,
as Mr. Paller will testify about today?
In terms of current federal efforts, I am particularly concerned
about what the National Cyber Security Division at DHS is doing right
now.
I am glad that the director of the NCSD is here today to answer
some of those questions. Mr. Purdy, for example, I also want to hear
more about what the NCSD is doing to help DHS complete the cyber
security portions of the National Infrastructure Protection Plan. A
final version of the NIPP was due last December. we are still waiting
for it.
I look forward to hearing from the members of this panel on all of
these issues.
Thank you Mr. Chairman.
Mr. Lungren. We are pleased to have a distinguished panel
of witnesses before us today on this important topic. The Chair
would recognize Mr. Donald ``Andy'' Purdy, the Acting Director
of the National Cyber Security Division of the U.S. Department
of Homeland Security, to testify.
I would just mention to all of you we are under the gun, I
am sorry about that, because of votes that we are going to
have. I would ask you to please restrict your oral statements
to 5-minutes, and your prepared statements will be made a part
of the record.
STATEMENT OF DONALD ``ANDY'' PURDY, ACTING DIRECTOR, NATIONAL
CYBER SECURITY DIVISION, U.S. DEPARTMENT OF HOMELAND SECURITY
Mr. Purdy. Good afternoon, Chairman Lungren and
distinguished members of the committee. My name is Andy Purdy.
I am the Acting Director of the Department of Homeland
Security's National Cyber Security Division. I am pleased to
appear before you today to share with you the work of NCSD to
address one of the significant threats to our cyberspace and
critical infrastructure, industrial control systems. In my
testimony today I will focus on our Control Systems Security
Program.
To carry out our mission and related responsibilities under
the National Infrastructure Protection Plan, we have identified
two overarching priorities: to build an effective national
cyberspace response system and implement a cyber risk
management program for critical infrastructure protection of
which our control systems effort is an important risk
mitigation effort.
The interdependency between physical and cyber
infrastructures is particularly acute in the use of control
systems as integral operating components by many of our
critical infrastructures. To assure immediate attention is
directed to protect these systems, we have established a
Control Systems Security Program to coordinate efforts among
Federal, State and local governments, as well as control
systems owners, operators and vendors, to improve control
system security within and across all critical infrastructure
sectors. As a key component of the program, in August, 2004, we
established a U.S. Computer Emergency Readiness Team Control
Systems Security Center in partnership with Idaho and Sandia
National Laboratories and other Department of Energy national
laboratories. The center's mission is to reduce the risk of
cyberattacks on control systems, and it partners with control
systems industry associations, universities, vendors and
industry experts.
Our program encompasses five goals. First we seek to
enhance the US-CERT capabilities for control systems security
to coordinate incident management, provide timely situational
awareness information, assess vulnerabilities, encourage
voluntary reporting and manage vulnerability and threat
reduction activities.
Our second goal is to reduce control system cyber
vulnerabilities in critical infrastructure. We have developed
the draft protection framework for identifying protection
measures and comparing them against existing security
standards. In addition, the framework includes a self-
assessment tool developed to allow owners and operators to
perform on-site assessments against the database of categorized
security requirements. We will soon pilot the tool with
multiple infrastructure sectors and will assist selected
control systems owners and operators in using the tool at their
sites.
Our third goal is to bridge industry and governmental
efforts through participation in working groups, standards
development bodies and user conferences. In partnership with
the Department of Homeland Security Science and Technology
Directorate, we chair the Process Control System Forum, which
includes industry, academia and government representatives. It
is designed to accelerate the development of technology that
will enhance the security, safety and reliability of control
systems, including legacy installations.
Our fourth goal is to develop control systems security
awareness and create a self-sustaining security culture within
the control systems community. A key element is our awareness
workshop program, which we began in May of this year and will
have completed approximately eight workshops by the end of this
year.
Our final goal is to make strategic recommendations for
improvements to future generation secure control systems and
security products. We have responsibility for developing
requirements for cybersecurity R&D projects to inform our
Science and Technology Directorate's research priorities, and
we coordinate with S&T in the development of new technologies
for securing control systems and networks.
We have a robust effort underway with our partners to
address the security of control systems through our Control
System Security Program. The efforts of our center toward
realizing the program goals has moved the ball forward in this
arena by increasing the control systems communities' awareness
of the need for cybersecurity and helping to provide them the
tools and resources to secure their control systems. We
continue to further these strategic goals through advancement
of our key initiatives.
We are committed to achieving success in meeting our goals
and objectives, but we recognize we cannot do it alone. We will
continue to meet and work with industry representatives, our
government counterparts, academia and State and local
government to formulate and enhance partnerships needed for
productive collaboration, and leverage the efforts of all so we
as a Nation are more secure in cyberspace and in our critical
infrastructure.
Again, thank you for the opportunity to testify to you
today, and I look forward to answering your questions.
Mr. Lungren. Thank you very much, Mr. Purdy.
[The statement of Mr. Purdy follows:]
Prepared Statement of Donald (Andy) Purdy, Jr.
Good morning Chairman King and distinguished members of the
Committee. My name is Andy Purdy, and I am the Acting Director of the
Department of Homeland Security's National Cyber Security Division
(NCSD). I am delighted to appear before you today to share with you the
work of the NCSD to address one of the significant threats to our
cyberspace and critical infrastructure--industrial control systems.
In my testimony today, I will provide an overview of NCSD's mission
and goals, priorities, and partnerships, with a particular focus on our
Control Systems Security Program. The Control Systems Security Program
addresses the cyber security of industrial control systems that run the
operational processes within the nation's critical infrastructure.
DHS and Critical Infrastructure Protection
Over the course of the past several months Secretary Chertoff
conducted a systematic evaluation of the Department's operations. On
July 13th, Secretary Chertoff announced the results of that evaluation
and outlined his six point agenda for the path ahead for the
Department. As part of this agenda, the Secretary announced several
Departmental organizational changes. Among these was the creation of a
new Preparedness Directorate which would house a newly created office
of the Assistant Secretary for Cyber Security and Telecommunications.
According to Secretary Chertoff, ``Securing our cyber systems is
critical not only to ensure a way of life to which we've grown
accustomed, but more importantly to protect the vast infrastructure
these systems support and operate.''
Currently, the Office of Infrastructure Protection (IP), located
within the Information Analysis and Infrastructure Protection (IAIP)
Directorate, is responsible for all critical infrastructure and key
resource protection. The Office of Infrastructure Protection has four
component divisions: (1) the Infrastructure Coordination Division
(ICD), (2) the Protective Security Division (PSD), (3) the National
Communications System (NCS), and (4) the National Cyber Security
Division (NCSD).
In December 2003, President Bush issued Homeland Security
Presidential Directive 7: Critical Infrastructure Identification,
Prioritization, and Protection (HSPD-7), which established a national
policy for federal departments and agencies to identify and prioritize
United States critical infrastructure and key resources and to protect
them from terrorist attacks. Among other things, HSPD-7 identified
seventeen (17) \1\ critical infrastructure and key resource sectors and
assigned responsibility for each to a Sector Specific Agency (SSA),
with DHS serving as the overall program coordinator.
---------------------------------------------------------------------------
\1\ The NIPP identifies the following Critical Infrastructure
Sectors and Key Resources: Food and Agriculture; Public Health and
Healthcare; Drinking Water and Wastewater; Energy; Banking and Finance;
National Monuments and Icons; Defense Industrial Base; Information
Technology; Telecommunications; Chemical; Transportation Systems;
Emergency Services; Postal and Shipping; Dams; Government Facilities;
Commercial Facilities; Nuclear Reactors, Materials, and Waste.
---------------------------------------------------------------------------
Additionally, HSPD-7 set forth how DHS should address critical
infrastructure protection, including development of a ``summary of
activities to be undertaken in order to: define and prioritize, reduce
the vulnerability of, and coordinate the protection of critical
infrastructure and key resources.''\2\ To meet this mandate, IP
developed the interim National Infrastructure Protection Plan (NIPP), a
plan that is to serve as the guide for addressing critical
infrastructure and key resource protection. It sets forth a risk
management framework for public and private sector stakeholders to work
together to identify, prioritize, and conduct vulnerability assessments
of critical assets and key resources in each sector. It also includes
the identification of interdependencies of critical assets and key
resources both within and across the sectors as well as providing
priority protective measures that owners and operators of such assets
should undertake to secure them. Recognizing that more that 85 percent
of the critical infrastructure is owned and operated by the private
sector and that the development of public-private partnership is
paramount to securing our nation's assets, private sector-led Sector
Coordinating Councils (SCCs) are being established to work with their
appropriate SSA via Government Coordinating Councils (GCC), which
represent the government agencies that have a role in protecting the
respective sectors.
---------------------------------------------------------------------------
\2\ Homeland Security Presidential Directive 7, December 17, 2003;
http://www.whitehouse.gov/news/releases/2003/12/20031217-5.html.
---------------------------------------------------------------------------
Currently, the Office of Infrastructure Protection is finalizing
the NIPP and it is expected to be released later this year. This
finalized document will refine the public-private partnership model and
a process for protecting our critical infrastructures from physical or
cyber attack or natural disasters.
DHS and Cyber Security
In June 2003, in response to the President's National Strategy to
Secure Cyberspace, the Department of Homeland Security created the NCSD
as a national focal point for cyber security. The national strategy
established the following five national priorities for securing
cyberspace:
Priority I:A National Cyberspace Security Response System
Priority II: A National Cyberspace Security Threat and Vulnerability
Reduction Program
Priority III: A National Cyberspace Security Awareness and Training
Program
Priority IV: Securing Government'S Cyberspace
Priority V: National Security and International Cyberspace Security
Cooperation
Given today's interconnected environment and DHS's integrated risk-
based approach to critical infrastructure protection, NCSD's mission is
to work collaboratively with public, private, and international
entities to secure cyberspace and America's cyber assets. To meet that
mission, NCSD developed a Strategic Plan that establishes a set of
goals with specific objectives for each goal, and milestones associated
with each objective. The Strategic Plan goals, which are closely
aligned with the Strategy, HSPD-7, the NIPP, and the Cyber Annex to the
recently announced National Response Plan, are as follows:
1. Establish a National Cyberspace Response System to prevent,
detect, respond to, and reconstitute rapidly after cyber
incidents;
2. Work with public and private sector representatives to
reduce vulnerabilities and minimize severity of cyber attacks;
3. Promote a comprehensive awareness plan to empower all
Americans to secure their own parts of cyberspace;
4. Foster adequate training and education programs to support
the Nation's cyber security needs;
5. Coordinate with the intelligence and law enforcement
communities to identify and reduce threats to cyberspace; and
6. Build a world class organization that aggressively advances
its cyber security mission and goals in partnership with its
public and private stakeholders.
To meet these goals, NCSD is organized into four operating branches
to address the various aspects of the risk management structure: (1)
U.S. Computer Emergency Readiness Team (US-CERT) Operations to manage
the 24x7 threat watch, warning, and response capability that can
identify emerging threats and vulnerabilities and coordinate responses
to major cyber incidents; (2) Strategic Initiatives to manage
activities to advance cyber security in critical infrastructure
protection, control systems security, software development, training
and education, exercises, and standards and best practices; (3)
Outreach and Awareness to manage outreach, cyber security awareness,
and partnership efforts to disseminate information to key
constituencies and build collaborative actions with key stakeholders;
and (4) Law Enforcement and Intelligence to coordinate with and share
information between these communities and NCSD's other constituents in
the private sector, public sector, academia, and others, and also to
coordinate DHS efforts within interagency response and mitigation of
cyber security incidents. Together, these branches make up NCSD's
framework to address the cyber security challenges across our key
stakeholder groups and build communications, collaboration, and
awareness to further our collective capabilities to detect, recognize,
attribute, respond to, mitigate, and reconstitute after cyber attacks.
The Strategy, HSPD-7, and the interim NIPP provide NCSD with a
clear operating mission and national coordination responsibility. To
carry out this mission and its related responsibilities, NCSD has
identified two overarching priorities: to build an effective national
cyberspace response system and to implement a cyber risk management
program for critical infrastructure protection. Our focus on these two
priorities and related programs addresses the overarching NIPP Risk
Management methodology and establishes the framework for securing
cyberspace today and a foundation for addressing cyber security for the
future.
Within the second priority, in addition to fulfilling our NIPP role
as the Sector Specific Agency for the Information Technology (IT)
Sector and providing cross-sector cyber security guidance to all
sectors, NCSD undertakes a cyber risk mitigation approach focused on
three key areas. These include the Internet Disruption Working Group,
the Software Assurance Program, and the Control Systems Security
Program.
NCSD and Control Systems Cyber Security
The interdependency between physical and cyber infrastructures is
hardly more acute than in the use of control systems as integral
operating components by many of our critical infrastructures. ``Control
Systems'' is a generic term applied to hardware, firmware,
communications, and software used to perform vital monitoring and
controlling functions of sensitive processes and enable automation of
physical systems. Specific types of control systems include Supervisory
Control and Data Acquisition (SCADA) systems, Process Control Systems
(PCS), and Distributed Control Systems (DCS).
Examples of the critical infrastructure processes and functions
that control systems monitor and control include energy transmission
and distribution, pipelines, water and pumping stations,
telecommunications, chemical processing, pharmaceutical production,
rail and light rail, manufacturing, and food production. Increasingly,
these control systems are implemented with remote access and
connections to open networks such as corporate intranets and the
Internet. Older control systems that operated with manual components,
vacuum actuators, and proprietary software are rapidly being upgraded
with modern computer systems. These sophisticated IT tools are making
our critical infrastructure assets more automated, more productive,
more efficient, and more innovative, but they also may expose many of
those physical assets to physical consequences from new, cyber-related
threats and vulnerabilities.
Control systems represent an attractive target for malicious actors
for several reasons. First, they provide a possible avenue for
inflicting physical, environmental, or economic harm to the nation from
a distance. Second, relatively mature attacking tools have been
developed and are available on the Internet. Finally, these tools can
be used with little technical expertise to attack control systems that
are accessible from the Internet.
To assure immediate attention is directed to protect these systems,
NCSD established the Control Systems Security Program to coordinate
efforts among federal, state, and local governments, as well as control
system owners, operators, and vendors to improve control system
security within and across all critical infrastructure sectors.
The Program incorporates five highly integrated goals to address
the issues and challenges associated with control systems security.
1. Coordinate control system incident management, provide
timely situational awareness information for control systems,
assess control system vulnerabilities, encourage voluntary
reporting, and manage control system vulnerability and threat
reduction activities by enhancing the US-CERT's capabilities
for control systems security;
2. Reduce control system cyber vulnerabilities in Critical
Infrastructure by establishing a proactive environment for risk
reduction and security assessments, to evaluate systems, and to
work with control systems owner/operators and vendors to
resolve vulnerabilities;
3. Bridge industry and governmental efforts through
participation in working groups, standards development bodies,
and user conferences to build cooperative and trusted
relationships and enhance control systems security efforts;
4. Develop control systems security awareness and create a
self-sustaining security culture within the control systems
community; and
5. Make strategic recommendations as to the funding,
development, and testing of next-generation secure control
systems and security products.
Goal 1--Enhance US-CERT capabilities for control systems cyber security
Our control systems activities support NCSD's overall efforts to
address cyber security across critical infrastructure sectors over the
long term, as well as the US-CERT's capability in the management,
response, and handling of incidents and vulnerabilities, and mitigation
of threat actions specific to critical control systems functions. NCSD
established the US-CERT Control Systems Security Center (CSSC) in
partnership with Idaho National Laboratory (INL) and other Department
of Energy (DOE) National Laboratories \3\ in August, 2004. Through the
use of Cooperative Research and Development Agreements (CRADA's) and
other mutually benefiting agreements, the CSSC also incorporates
partners from control systems industry associations, universities,
vendors, and industry experts. The CSSC mission is to reduce the risk
of cyber attacks on control systems, and as such, it provides
facilities and expertise to support the reduction of risk in critical
infrastructure through site and system assessments, demonstrations for
education and awareness, risk assessment and risk analysis, adversarial
awareness, and coordination among the national laboratories.
---------------------------------------------------------------------------
\1\ Pacific Northwest, Los Alamos, Argonne, Sandia, Lawrence
Livermore and Savannah River
---------------------------------------------------------------------------
Through its partnerships and technological improvement efforts for
systems and facilities, the CSSC has been maturing response
capabilities to support US-CERT with control system expertise. The CSSC
continues to work with the US-CERT in enhancing their ability to
provide initial control system guidance and expertise, and a CSSC
limited access secure portal (https://us-cert.esportals.net/) has been
established for information coordination and dissemination of cyber
threat and vulnerability alerts. A web site is under development to
share control systems security information with our cyber security
partners and the control systems community. The web site, which will be
available in FY06, will also provide information, resources, and links
for owners and operators to effectively defend their control systems. A
``Tier II'' support function will further support US-CERT by leveraging
CSSC partners in incident response and vulnerability handling, and
performing in-depth evaluation of specific attacks or exploits and
determining the impact on various operating systems, components, and
vendor systems.
In FY06, CSSC will explore the need for establishing a trusted
third-party within academia to serve as a voluntary reporting center to
encourage open communication among the private sector regarding
emerging control system threats and exploits. As such, the CSSC is
developing a control systems incident management support tool to
enhance US-CERT cyber threat notification efforts. It is designed for
use when a new vulnerability is detected and will enable the
identification of critical infrastructure at greatest risk to an
identified threat, thereby enabling the CSSC to rapidly notify the
facilities at the greatest risk. Owners and operators can then
implement protective measures as appropriate to reduce that risk and
mitigate damage to their systems. It is important to note that the
effectiveness of the tool is dependent on the acquisition of current
owner/operator system data. NCSD continues to work with Sector Specific
Agencies to obtain data from the various sectors necessary to utilize
the tool and maximize its benefits.
Goal 2--Reduce control system vulnerabilities in critical
infrastructure
To reduce control system vulnerabilities in our critical
infrastructure, CSSC developed a draft cyber security protection
framework for identifying control systems security protection measures
and comparing them against existing security standards. The cyber
security protection framework, which is based on the Common Criteria
and an Industrial Control System Security Protection Profile developed
by the National Institute of Standards and Technology, supports NCSD's
mission to reduce cyber security risk within control systems. The
framework provides a systematic methodology for assessing the cyber
security posture of control systems. It is designed to reduce the
burden on owners and operators by providing them with a means to select
protective measures that apply to their specific architecture and
operating environment and reduce their respective risk.
Application of the framework methodology results in a risk-based
set of security measures. Risk is defined by DHS as Risk = Threat x
Vulnerability x Consequence. To calculate quantitative values for risk,
one must define the system of interest, establish attack-defense-
failure scenarios, and consider the consequences of a successful
attack. Then, protection measures are identified to reduce risk. The
overall goal is to provide a quantitative, traceable, and supportable
value of risk.
As part of this framework, the CSSC also has capabilities at INL to
perform vulnerability assessments of control systems. For example, the
CSSC leverages the National SCADA Test Bed funded by DOE and operated
in partnership with Sandia National Laboratories. Linkages with these
test beds and assessment facilities provides the CSSC with incoming and
outgoing data traffic and communication channels necessary for the
replication of control systems (e.g., PCS, SCADA) and components. These
testing capabilities also support quick mock-ups of control systems
and/or components to evaluate existing threats, vulnerabilities, and
incidents as they are reported to the US-CERT.
The CSSC utilizes a unique ``plug and play'' patching system that
allows engineers to assess systems or components in an environment
simulating the conditions found in industry to include multiple
communication pathways and live incoming and outgoing control systems
specific data traffic. This allows for in-depth assessments of control
systems in a near true-to-life environment. The CSSC is working with
commercial vendors and DOE to complete assessments of three different
control systems to identify cyber vulnerabilities, reverse engineer
exploits, and provide solutions to secure vendor systems. A code-based
analysis has also been conducted in cooperation with a vendor/
manufacturer to identify possible vulnerabilities and recommendations
to secure the system.
Our adversaries are developing tools to hack into and take over
control systems, and we need greater collective awareness of those
capabilities to understand specific threats to and vulnerabilities of
our control systems. As such, CSSC tracks information on current
control systems security trends and threats, review and assesses new
vulnerabilities and exploits as they are discovered or reported, and
conducts analysis to better understand adversarial tools and
capabilities. The CSSC considers specific exploit assessment scenarios
on control systems and ``reverse engineers'' exploits to provide
solutions to industry before an exploit is made public.
The cyber security protection framework also leverages best
practices from industry for securing control systems against cyber
attacks and organizes them so the control systems community can
identify specific solutions to their security vulnerabilities. As part
of the framework, implementation tools, such as a ``self-assessment
tool,'' have also been developed to allow owners and operators of
industrial control systems to perform on-site self-assessments against
a database of categorized security requirements. Each security
requirement is supported by recommendations for meeting the requirement
and mitigating vulnerabilities within the architecture of that
particular control system. As new vulnerabilities emerge and associated
solutions are developed, the framework of security requirements will
expand and new protection solutions will be made available to the
control system community. The protection framework provides categorized
and graded guidance, component by component, for improving cyber
security of control systems.
The draft security protection framework and its associated
implementation tools are ready for validation. NCSD will soon pilot the
self-assessment tool with multiple infrastructure sectors and will
assist selected control system owners and operators in using the tool
at their sites. This effort will help owners and operators identify
security vulnerabilities within their systems, recommend solutions for
reducing the risk of successful cyber attacks, and prioritize risk
reduction efforts. The pilot effort will also allow NCSD to validate
and enhance the self-assessment tool for future, widespread roll-out
across the control system community. NCSD is also working with PSD and
other Sector Specific Agencies to ensure that concepts from the cyber
security protection framework are integrated into risk and
vulnerability assessments across the sectors. For example, NCSD is
working closely with the American Society of Mechanical Engineers and
PSD to incorporate cyber into the Risk Analysis and Management for
Critical Asset Protection (RAMCAP) framework.
Goal 3--Bridge industry and governmental efforts through
participation in working groups, standards development bodies, and user
conferences
A primary objective of NCSD's Control Systems Security Program is
to coordinate efforts among Federal, State, and local governments, as
well as control system owners, operators, and vendors to improve
control systems security within and across all critical infrastructure
sectors.
In partnership with DHS' Science and Technology (S&T) Directorate,
NCSD chairs the Process Control System Forum (PCSF). The PCSF includes
industry, academia, and government representatives and is designed to
accelerate the development of technology that will enhance the
security, safety, and reliability of control systems, including legacy
installations.
In addition to the PCSF, the CSSC works to enhance private sector
awareness through participation in industry association meetings, user
groups, and standards coordination work groups. For example, most
recently, representatives from CSSC participated in a Railroad
Association meeting in Annapolis, Maryland, the Pacific Northwest
Economic Region 15th Summit, and the Interagency Forum for
Infrastructure Protection in Portland, Washington. At all of these
gatherings, attendees were provided with an overview of the CSSC
program, capabilities, and with information on how they can participate
and take advantage of what the CSSC program has to offer, including
alert and informational bulletins, self-assessment and risk reduction
calculation tools.
CSSC has also established relationships with a number of industry
partners, including partnerships designed to facilitate initial
assessments and develop risk reduction plans in various industry
sectors. Our private industry partners provide experience in
understanding vulnerabilities and operational perspectives, and bring
established contacts within the control systems community.
Specifically, they provide CSSC with control system expertise from
various critical infrastructure sector perspectives; expertise and
feedback on assessment tools; subject matter expertise regarding
development of security requirements and best practices; assessment,
research, and risk assessment capabilities; and contacts and
opportunities to interface with sectors.
CSSC is also working with control system vendors to provide
equipment for assessments to be conducted at CSSC facilities. They
assist in identifying vulnerabilities based on their experience and
work to resolve vulnerabilities in next generation and legacy systems
as a result of assessments performed against their systems. A number of
industries (e.g., oil and gas, chemical, petro-chemical, electrical,
power generation plant automation [coal, hydro, and gas fired plants],
and transportation) are contributing to these CSSC efforts to reduce
cyber vulnerabilities in control systems. Partnerships with members of
the control system community are designed to help NCSD better assist
owners and operators secure their systems.
Goal 4--Enhance control systems security awareness
The NCSD is engaged in several activities designed to increases
awareness and provide the tools and products necessary to enable the
critical infrastructures and key resources to secure their control
systems against cyber threats. A key element is CSSC's awareness
workshop program.
Our ``threat-brief, demonstration, and mitigations'' workshop has
been well received by the control systems community. The first workshop
was held in May, 2005 at a PCSF meeting in Dallas, Texas. Since then
additional workshops have been held in Bellevue, Washington and Idaho
Falls, Idaho. We anticipate that by late 2005, approximately eight
workshops will have been conducted. The workshops include a brief
overview of the threat picture, a cyber vulnerability demonstration,
and a discussion of mitigation steps. NCSD has found that cyber
vulnerability demonstrations are an effective method to show the impact
that cyber attacks can have on their control systems and operations and
that cyber security is essential to protect them.
Goal 5--Make strategic recommendations for improvements to future
generation secure control systems and security products
Cyber-related research and development (R&D) is vital to improving
the resiliency of the Nation's critical infrastructures. This difficult
strategic challenge requires a coordinated and focused effort from
across the Federal Government, State and local governments, the private
sector, and academia to advance the security of critical cyber systems.
Two components within DHS share responsibility for cyber R&D. The
Science & Technology (S&T) Directorate serves as the primary agent
responsible for executing cyber security R&D programs. NCSD has
responsibility for developing requirements for cyber security R&D
projects. NCSD supports the overall DHS R&D mission by identifying
areas for cyber innovation and coordinating with S&T. NCSD collects,
develops, and submits cyber security R&D requirements to provide input
to the federal cyber security R&D community and specifically to inform
the DHS S&T Directorate's cyber security research priorities. NCSD
coordinates with S&T on the development of new technologies for
securing SCADA systems and networks.
NCSD's Control Systems Security Program identifies R&D cyber
security requirements for legacy and next generation control systems
and security products through US-CERT CSSC operational activities such
as incident management, site and system assessments, and analyses. As
difficult problems which would benefit from advanced technological
solutions are discovered, requirements are identified and forwarded to
control systems vendors and DHS S&T for new R&D projects. Best
practices, common vulnerabilities, and requirements for security
standards are also shared with the control systems community to promote
enhanced security for legacy and new control systems.
DHS S&T manages the Congressionally directed funding for the
Institute for Information Infrastructure Protection (I3P). The I3P is a
national research consortium composed of more than two dozen research
entities, including academic institutions, non-profits, federally
funded labs, and FFRDCs. In early 2005, the I3P launched a major
initiative focused on addressing the vulnerabilities of SCADA systems
in the oil and gas industry.
Moving Forward
NCSD has a robust effort underway to address the security of
control systems through our Control Systems Security Program. The
efforts of the CSSC toward realizing the five goals the Program sets
forth, including the enhancement of capabilities, initiatives to reduce
vulnerabilities, and establishment of partnerships, has moved the ball
forward in this arena by increasing the control system communities'
awareness of the need for control systems cyber security and providing
them the tools and resources to secure their control systems.
Many activities are planned for the near future including:
Developing and finalizing the CSSC portal and web site
to enhance capabilities and encourage greater information
exchange with the control system community.
Supporting vulnerability assessments to determine the
cyber security posture of legacy and next generation control
systems at critical sites. Assessments will identify critical
components threat vectors, and misconfigurations in hardware,
applications, and network topologies within our current
infrastructure and recommend protective measures. This
information will aid in determining the level of compliance
with current best practices and control system protection
framework requirements.
Continuing to integrate CSSC activities, skills, and
capabilities to identify particular high risk cyber
vulnerabilities. Specifically, for FY06 high-risk system
vulnerabilities will be identified in at least two critical
infrastructure sectors and then security enhancements to
mitigate those vulnerabilities will be identified. Other site
assessments will be supported as appropriate to identify cyber
risks to control systems.
Encouraging the voluntary implementation of security
measures. The CSSC will accomplish this through development of
a ``Business Case,'' beginning in FY06. Development of a
business case will demonstrate cost-benefit where the cost will
be represented as the cost of implementing countermeasures and
benefit will be the reduction of risk. Risk analysis is the
basis for the business case.
Continuing to work with PSD and other Sector Specific
Agencies to integrate cyber security and control systems
security efforts into risk and vulnerability assessment efforts
such as Comprehensive Reviews, the Vulnerability Identification
Self Assessment Tool, and the Risk Analysis and Management for
Critical Asset Protection.
Continuing to participate in forums and meetings to
raise awareness while conducting targeted outreach activities
in sectors and with senior executives to not only pilot and
validate our control systems protection framework and tools but
also to create an understanding among control system owners and
operators of the need for and importance of security.
We are committed to achieving success in meeting our goals and
objectives, but we cannot do it alone. We will continue to meet with
industry representatives, our government counterparts, academia, and
state and local representatives to formulate the partnerships needed
for productive collaboration and leverage the efforts of all, so we, as
a nation, are more secure in cyberspace and in our critical
infrastructures.
Again, thank you for the opportunity to testify before you today. I
would be happy to answer any questions you have.
Mr. Lungren. The Chair would now recognize Mr. Larry Todd,
the Director of Security, Safety and Law Enforcement for the
Bureau of Reclamation, U.S. Department of Interior, to testify.
STATEMENT OF LARRY TODD, DIRECTOR OF SECURITY, SAFETY AND LAW
ENFORCEMENT, BUREAU OF RECLAMATION, U.S. DEPARTMENT OF THE
INTERIOR
Mr. Todd. Thank you, Mr. Chairman and distinguished members
of the subcommittee. I am pleased to appear before you today to
tell you about the security of the control systems used by the
Bureau of Reclamation.
Reclamation uses SCADA systems as tools to enable us to
meet our mission of water delivery, power generation, flow
monitoring and water regulation. SCADA is used to control
outlet works, valves at dams, to control hydroelectric
generators and associated circuit breaker switches and
transformers, and to control pumps and gates on water delivery
systems and canals. However, we do not use SCADA controls to
operate the spillway gates, nor for flood control operations.
Reclamation has a number of security features built into
the SCADA operation. For instance, no SCADA system is attached
to the Internet, and therefore, the systems cannot be accessed
by the Internet. There are software controls within SCADA
systems to protect against unauthorized operation, and on some
facilities we have mechanical controls that prevent operation
beyond set parameters. In addition, Reclamation regularly tests
to ensure the connectivity does not exist.
To help identify physical and cyber vulnerabilities,
Reclamation uses independent organizations to evaluate our
security posture. We have had numerous investigations by the
Inspector General's Office, and they report that the SCADA
systems are operating in relative safety from potential
catastrophic cybersecurity threats.
In summary, Reclamation recognizes that SCADA plays a key
role in protecting critical infrastructure components. Where we
employ SCADA systems, we believe we have taken responsible
steps to ensure their security and safe operation. We also will
employ better assessment and protection tools as they become
available.
Thank you for this opportunity to describe Reclamation's
use of SCADA. I would be pleased to answer any questions the
committee may have.
Mr. Lungren. Thank you very much, Mr. Todd.
[The statement of Mr. Todd follows:]
Prepared Statement of Larry Todd
Mr. Chairman, my name is Larry Todd, and until recently I served as
the Director of Security, Safety, and Law Enforcement for the U.S.
Bureau of Reclamation. Established in 1902, Reclamation is known
primarily for the dams, power plants, and canals we have built and
operate in seventeen western States. Reclamation is our Nation's
largest wholesaler of water, and its second largest producer of
hydroelectric power. I am pleased to appear before you today to tell
you about the security of the control systems used by the Bureau of
Reclamation.
Reclamation's Supervisory Control and Data Acquisition (SCADA) Systems
Reclamation employs SCADA systems as tools to enable us to meet our
mission obligations of providing essential services and commodities.
These obligations include electric power generation, flood monitoring,
water regulation, and water delivery. To accomplish these goals,
Reclamation controls water release gates and valves at dams;
hydroelectric generators, circuit breakers, switches and transformers
at power plants; and pumps and gates on waterways and canals.
Reclamation's SCADA systems collect information about our
facilities through transducers, converting information such as gate
position, reservoir level, hydroelectric generator output, and water
flow to electrical signals for processing in the SCADA system's
computers. Once in the computers, the information is examined for any
unusual characteristics, such as whether it exceeds an expected value.
When information does not meet expectations, alarms may be triggered to
inform operations staff of the situation, enabling them to take
corrective actions. Reclamation's major SCADA control centers are
manned at all times, enabling operations staff to react to both normal
operations and emergency situations 24 hours a day and 365 days a year.
Along with collecting information, Reclamation's SCADA systems also
facilitate our operations staff's reaction to normal and abnormal
operational needs. They do this by supporting the supervised remote
control of our facilities. By providing the operations staff with
information about the facility, informed decisions can be made quickly
and the appropriate actions taken. The SCADA systems computers help to
supervise these decisions by ensuring that they meet safe operational
criteria.
Protecting Reclamation's SCADA Systems
The focus of security efforts has changed since SCADA systems were
first employed by Reclamation. In those early years SCADA design
focused almost entirely on the operational integrity of the SCADA
systems. In all cases where SCADA systems were permitted to control
equipment, the safety and reliability of the control was examined and
appropriate improvement measures were engineered and incorporated. This
supported safer equipment operation and permitted the disabling of
SCADA control if necessary. This was done to protect the equipment and
to ensure the safety of the public and Reclamation personnel in the
event of a SCADA malfunction. These safety measures acted independently
from the SCADA system to ensure that the failure of the SCADA system
did not adversely affect the safety measures. If the safety of SCADA
control actions could not be ensured, additional steps were taken to
limit the degree of SCADA control or the control was not enabled.
Reclamation still follows these practices in implementing its SCADA
systems, providing a significant measure of operation security for its
SCADA controlled facilities.
From the very beginning of Reclamation's use of SCADA systems, we
have maintained a policy of not connecting our SCADA systems to our
administrative networks. Today we adhere to that policy in all but the
most unusual of situations. All connections to SCADA systems are
minimized. Reclamation does not connect its SCADA systems to the
Internet and routinely tests to ensure that such connectivity does not
exist. Wherever practical, connections to our SCADA systems do not use
Internet-like protocols, instead employing simple, limited capability,
serial protocols. Those connections that must be present and that use
Internet-like protocols are protected by firewalls and intrusion
detection systems. Reclamation has adopted ``best practices'' and
follows the cyber security guidance outlined by the National Institute
of Standards and Technology (NIST) in their Special Publications.
In addition, Reclamation has evaluated and improved both personnel
and physical security at our SCADA facilities. We perform background
checks on key personnel and have ``hardened'' our facilities and
control rooms through the addition of various access controls. This
includes the access to our SCADA system control consoles.
To help identify physical and cyber vulnerabilities within the
organization, Reclamation has invited independent organizations,
including some represented by other panel members, to evaluate our
security posture. We have also supported numerous investigations by our
Inspector General's Office, some of which included limited penetration
testing of our SCADA systems. The Inspector General's FY05 management
report concluded that ``the SCADA systems are operating in relative
safety from potentially catastrophic cyber-security threats.'' To
maintain these results, we are continuously evaluating and implementing
prudent and practical security improvements.
Actions to Improve SCADA Security
Despite our security successes so far, Reclamation believes we can
still take additional steps to improve the security of our SCADA
systems. These steps, specifically identified and addressed in internal
documents, will create more rigorous testing processes, improve and
increase the frequency of security assurance reviews, and establish
more comprehensive security planning targets. We also favor additional
steps to improve the coordination of SCADA security efforts at both the
Federal and private sector levels. Close coordination will assure
consistency of Federal and private sector standards and security
guidance, and could also help ensure that an appropriately rigorous
security baseline is established for SCADA systems employed in
different industry segments, depending on the significance of the
infrastructure monitored or controlled.
In Summary
Reclamation recognizes that it plays a key role in protecting
critical infrastructure components, including dams, waterways, water
resources, and electrical generation capability. Where we employ SCADA
systems to facilitate the control of these components, we believe we
have taken responsible steps to ensure their security and safe
operation. We recognize that cyber security, as it applies to both
administrative and SCADA systems, requires continuous monitoring and
diligence. We believe our security program meets the challenges of
these requirements, but look forward to contributing to and employing
better development, assessment, and protection tools and techniques as
they become available.
Mr. Lungren. The Chair would now recognize Mr. Sam Varnado,
the Director of Information Operations Center at the Sandia
National Laboratory, to testify.
STATEMENT OF SAM VARNADO, DIRECTOR, INFORMATION OPERATIONS
CENTER, SANDIA NATIONAL LABORATORY
Mr. Varnado. Thank you, Mr. Chairman and distinguished
members of this committee. I am Sam Varnado from Sandia
National Laboratories, with laboratories in both California and
New Mexico.
First let me applaud the work the committee is doing. It is
very important to the well-being of our citizens and to the
national security. I am pleased to be part of it.
Today we are going to discuss SCADA systems. We are
concerned about these systems. We are very worried about them
because successful cyberattacks on these systems could lead to
serious consequences, which include loss of life, destruction
of equipment that is hard to replace, environmental insult and
economic loss.
Let me give you one example. Mr. Chairman, in June of 1982,
a huge explosion occurred in the Siberian wilderness in the
former Soviet Union. The yield was estimated at 3 kilotons in
that explosion. In his book At the Abyss: An Insider's History
of the Cold War, Thomas Reed attributes the monumental
explosion and resulting fire to a cyberattack on the SCADA
system that controlled the Trans-Siberian pipeline. According
to Mr. Reed, the pipeline software that ran the pumps, turbines
and valve settings was programmed to produce pressures far
beyond those acceptable to the pipeline joints and wells. He
further states that the malevolent software in this case was
what we call today a Trojan. It had been implanted in the host
software by a foreign intelligence service. This episode
illustrates the physical damage that can be created by
attacking a cybersystem.
SCADA systems are the soft underbelly of our infrastructure
protection strategy in this country. The older stand-alone
legacy SCADA systems are highly vulnerable. Some of these
vulnerabilities are listed in my written statement. But today
the trend is to replace those older systems with control
systems that use the Internet as the backbone. From a security
standpoint, this will make matters worse for the following
reasons: First, U.S. computer networks are under daily attack,
and adversaries are becoming more sophisticated. We are seeing
structured, well-resourced attacks that are designed to steal
information or disrupt and/or deny processes. For example, the
recent Super Slammer, which was a fast worm, infected 60
percent of DOD's NIPRNet computers in 8 minutes.
Improvements in attack methods, particularly by
sophisticated threats such as terrorist and nation states, are
outpacing our activities in defensive countermeasures. The
contest between the attackers and the defenders is a dreadful
mismatch with the advantage strongly in the attacker's corner.
Second, information technology vendors release on average
four new vulnerabilities each day at the same time new attack
methods are proliferating.
Third, we have no alternative to the use of commercial off-
the-shelf, or COTS, products in our information systems because
of cost issues; therefore, most of the hardware and software we
use is manufactured in countries whose interests do not always
align with those of the United States. We are buying and
embedding these products in very complex systems that we expect
to be secure. We are essentially trying to build trusted
systems from untrusted components, and many of us wonder if it
can be done at all.
Fourth, most of the current emphasis in cybersecurity is on
responding to hacker attacks that exploit the inherent
vulnerabilities that are present in all networked computer
systems. This effort is necessary and useful and should be
increased, but a longer-term view is needed. We need to put
more emphasis on addressing enterprisewide solutions and
threats from the more sophisticated adversaries.
My suggestions for addressing these problems are as
follows: First, reaffirm the concept of public/private
partnerships, and encourage stronger collaboration among
government, industry, universities and national labs. We need
to put more effort into sharing information on threats,
vulnerabilities, consequences of outages, training and
technology.
Second, extend these partnerships to include helping the
infrastructure owners make the business case for their
investments in security upgrades.
Third, increase funding for cybersecurity technology to
address the new threat and vulnerability environment and to
keep the defensive efforts on par with the attack development
activities being conducted by our adversaries.
Fourth, establish and fully fund a concentrated effort to
provide defense against the sophisticated threat.
Finally, support the initiatives, directives and plans
described in several reports that DHS and the administration
have produced over the last few years.
Thank you, Mr. Chairman and members of the committee, for
the opportunity to address you today. I would be happy to
answer questions at the appropriate time.
Mr. Lungren. Thank you, Doctor.
[The statement of Mr. Varnado follows:]
Prepared Statement of Dr. Samuel G. Varnado
Introduction
Mr. Chairman and distinguished members of the committee, thank you
for the opportunity to testify on the vulnerabilities of, and threats
to, Supervisory Control and Data Acquisition (SCADA) systems. I am Dr.
Sam Varnado, Director of Sandia National Laboratories' Information
Operations Center. I have more than thirty years of experience in
energy, information, and infrastructure systems development. I
currently coordinate Sandia's activities in cyber security technology
development, with special emphasis on critical infrastructure
protection applications.
Sandia National Laboratories is managed and operated for the
National Nuclear Security Administration (NNSA) of the U.S. Department
of Energy (DOE) by Sandia Corporation, a subsidiary of the Lockheed
Martin Corporation. Sandia's unique role in the nation's nuclear
weapons program is the design, development, qualification, and
certification of nearly all the nonnuclear subsystems of nuclear
warheads. We perform substantial work in programs closely related to
nuclear weapons--including intelligence, non-proliferation, and treaty
verification technologies. As a multiprogram national laboratory,
Sandia also conducts research and development for other federal
agencies when our special capabilities can make significant
contributions.
My statement will describe SCADA systems, identify some of the
threats they face, describe some of the cyber vulnerabilities of these
systems, discuss the consequences of disruptions, and explain Sandia's
contributions and capabilities in SCADA system security. I will also
comment on the gaps in current approaches to the problem, possible
solutions, and needs that Congress might choose to address.
What Are SCADA Systems and How Are They Used in Critical Infrastructure
Applications?
Both the national security of the United States and the well being
of our citizens are highly dependent on the reliable operation of the
nation's critical infrastructures. These infrastructures include
electric power, oil and gas, banking and finance, transportation,
telecommunications, and other networks. The operation of most of these
infrastructures is controlled by SCADA systems. These systems are
highly vulnerable to a wide range of threats, including terrorism. As
an example, we have shown that it is possible to turn out the lights in
most major U.S. cities through cyber attacks on SCADA systems.
Disruption of these systems by any means will result in substantial
economic loss, potential loss of life, long recovery times, and severe
disruption of the lives of our citizens.
We should note that we use the term ``SCADA'' to include all real-
time digital control systems, process control systems, and other
related technologies. The control processes for each infrastructure are
automated systems that combine humans, computers, communications, and
procedures. Automated systems are used to increase the efficiency of
process control by replacing high-cost personnel with lower cost
computer systems. The widespread use of SCADA systems makes them
critical to the safe, reliable, and efficient operation of physical
processes common to most infrastructures.
High Level SCADA Vulnerabilities
SCADA systems have generally been designed and installed with
little attention to security. Terrorist groups are aware of this. As
noted in an article in the June 27, 2002 Washington Post, these systems
have been targeted by al-Qa'ida terrorists. Some government experts
have concluded that the terrorists hope to use the Internet as an
instrument of bloodshed by attacking the juncture of cyber systems and
the physical systems they control. The article further postulated that
combined cyber and physical attacks could produce nightmarish
consequences.
Sandia has been investigating vulnerabilities in SCADA systems for
over ten years. During this time, many have been found. Our red team
assessments show that security implementations are, in many cases,
nonexistent or poorly implemented. Many of the older SCADA systems are
operated in a stand-alone mode; that is, they are not connected to the
Internet or to other corporate systems. Even so, these legacy systems
have vulnerabilities, including inadequate password policies and
security administration, no data protection mechanisms, and information
links that are prone to snooping, interruption, and interception. When
firewalls are used, they are sometimes not adequately configured, and
there is often a ``back-door'' access because of connections to third-
party contractors and maintenance staff. We have found many cases in
which unprotected remote access allows users to circumvent the
firewall. In addition, most of the SCADA manufacturers are foreign-
owned.
In summary, it is easy for adversaries to take control of these
legacy systems and cause disruptions with significant consequences.
Today, the legacy systems are gradually being replaced by new SCADA
systems that use the Internet as the control backbone. This change is
being implemented to reduce cost and increase efficiency of operation.
However, this trend substantially increases the possibility of
disruptions because (1) the number of people having access to the
system is substantially increased, (2) disruptions can be caused by
hackers who have no training in control systems engineering, and (3)
the use of the Internet exposes SCADA systems to all the inherent
vulnerabilities of interconnected computer networks that are currently
being exploited by hackers, organized crime, terrorists organizations,
and nation states. Worms, viruses, network flooding, no-notice attacks
through compromised routers, spyware, insider attacks, data
exfiltration by outsiders who gain insider privileges (phishing), and
Distributed Denial of Service attacks are all commonplace. Effectively
combating these attacks requires increased awareness, new technology,
and improved response and recovery capabilities.
Especially vulnerable is the electric power grid. Under
restructuring, the grid is now being operated in a way for which it was
never designed. More access to control systems is being granted to more
users, the demand for real-time control has increased system
complexity, and business and control systems are interconnected. In
many cases, these new systems are not designed with security in mind.
More vulnerabilities are being found, and the opportunities for
disruptions are increasing rapidly. The complexity of the systems and
the high degree of interdependency among the infrastructure sectors can
lead to cascading failures in which failures in one sector can
propagate to others.
Sandia has identified the vulnerabilities of SCADA systems and
summarized them in a report--''Common Vulnerabilities in Critical
Infrastructure Control Systems''--that is available from our Center for
SCADA Security website (http://www.sandia.gov/scada). The report
identifies the vulnerabilities that we uncovered in our red team
assessments of systems in use by a diverse set of customers from the
electric power, petroleum, natural gas, and water infrastructures. This
document has been made available to other government agencies and to
private industry.
SCADA Threats
Sandia performs vulnerability assessments using a red team process
that models adversarial capabilities and approaches. It is essential to
view SCADA systems from an adversarial perspective in order to identify
their important vulnerabilities. We use adversarial modeling as a way
of understanding threats from different political, social, and
motivational structures so that relevant characteristics may be
utilized to identify the classes of attacks that each adversary might
be able to launch. Hackers, organized crime, cyber terrorists, and
nation states are examples of different classes of adversaries with
varying capabilities and attributes.
We consider two basic categories of adversaries: ``outsiders'' and
"insiders." It is generally the goal of an outsider to acquire the
attributes of an insider through such means as hijacking connections,
password sniffing, and identity theft. Most U.S. critical
infrastructure owners and operators have only a passing knowledge of
the nature of the adversaries' capabilities. Consequently, the level of
protection is low and the probability of significant disruptions is
high. Critical infrastructure owners and operators need to increase
their awareness of both the vulnerabilities and the threat. They also
need training in network defense, information about improvements in
cyber security technology for control systems, and timely updates on
threat information.
SCADA Attack Consequences
The consequences of disruptions to SCADA systems are numerous,
expensive, and varied. Two examples are presented here simply to make
the point that we must start thinking seriously about the security of
SCADA systems.
In his book, At the Abyss: An Insider's History of the Cold War,
Thomas C. Reed (former National Security Council member and Air Force
Secretary) reported that in June 1982 the CIA, through exploitation of
software transferred to the Soviet Union, created a damaging attack on
Soviet pipeline systems. The software that was used to run the pumps,
turbines, and valves of the pipeline was programmed to malfunction
after a specific time interval. The malfunction caused the control
system to reset the pump speeds and valve settings to produce pressures
beyond the failure ratings of the pipeline joints and welds. The result
was the largest non-nuclear explosion and fire ever seen from space.
There were no physical casualties, but the goal of economic damage was
met. This story is an excellent example of the type of attack that can
be accomplished by a nation state.
In January 2003, when the SQL Slammer worm began attacking computer
networks around the world, users of the business network at Ohio's
Davis-Besse nuclear power plant began to notice a network slowdown.
Investigation revealed the worm had spread from the plant's business
network to its operations network, causing enough congestion to crash
the computerized panel used to monitor the plant's most crucial safety
indicators. Minutes later, the Plant Process Computer, another
monitoring system, crashed as well. The plant's firewall had initially
blocked Slammer, but the worm still managed to reach the plant through
a high-speed connection from an unsecured contractor's network. Had the
plant's operations network been properly protected from either the
contractor's network or the plant's own business network--or had the
plant operators installed Microsoft's patch to prevent the Slammer
infection (released six months earlier)--the infiltration would not
have happened. Fortunately, the incident did not result in disaster
because the plant was off-line at the time, for regular maintenance,
and the crashed monitors were being backed-up by analog counterparts..
These two incidents exemplify the potential consequences of
inadequate cyber security processes. We should regard them as warnings.
Sandia's Contributions to Critical Infrastructure Control System
Protection
SCADA Security and Standards
During the Clinton administration, Sandia was heavily involved in
supporting the President's Commission on Critical Infrastructure
Protection. That activity, along with our experience in providing
secure information systems for nuclear weapon command and control
systems, provided impetus for our initial work in SCADA security. We
began our work with laboratory directed research and development (LDRD)
funds, and we initiated development of a laboratory SCADA test bed in
1998. At that time it was difficult to convince others of the
implications of SCADA vulnerabilities, so we also engaged the standards
community. Standards are necessary for improving the security of
distributed, networked systems. Because many SCADA equipment
manufacturers are foreign owned, the only way to provide trusted
systems is through the application of standards. Sandia was designated
by the DOE to be the U.S. representative to the International
Electromechanical Committee standards working group, TC57. We are
expanding our efforts, in collaboration with other national
laboratories, by engaging other standards groups like AGA 12-1
(``Cryptographic Protection of SCADA Communications''), API 1164 (``API
Security Guidelines for the Petroleum Industry''), and ISA SP99
(``Manufacturing and Control System Security''), as well as various
IEEE working groups.
Sandia maintains strong research and development programs in
cryptography, network security, secure network architecture design,
wireless network security, threat assessment, and intelligent agent-
based security approaches. This work is coordinated by our Center for
SCADA Security, which was established in 2000.
Red Team and Assessments
Sandia also performs vulnerability assessments of critical
infrastructure systems from both cyber and physical security
perspectives. We have completed vulnerability assessments of a number
of dams in the western United States. We have also assessed the
vulnerability of networks used by a number of banks and by the
Strategic Petroleum Reserve. We have worked with the electricity and
oil and gas sectors to improve the robustness of their SCADA systems.
As a result of these experiences--as well as our own strategic
planning, our LDRD investments, and the foresight of sponsors to invest
resources toward critical infrastructure protection--Sandia was in a
position to immediately address some of the urgent needs following the
events of 9/11.
For example, we quickly developed a self-assessment methodology
called RAM-W for water treatment facilities; this effort was sponsored
by both the Environmental Protection Agency and the American Water
Works Association Research Foundation. We also developed training
classes on assessing SCADA systems for use in training our own staff.
We now provide this training to industry, and we promulgate best
practices to industry for securing SCADA systems. These and other
contributions to critical infrastructure protection are possible
because of strategic planning conducted years ago that led to early
investment in the capabilities needed to respond. We also continue to
invest LDRD funds in areas of urgent need. Examples include the
integration of cyber and physical security technology, cryptographic
solutions for SCADA system communications, modeling and simulation of
infrastructure elements, secure control of micro-grids, SCADA
forensics, and application of new network security technologies to
SCADA systems.
Partnering Activities
In 2004, the DOE and the National Energy Technology Laboratory
funded the National SCADA Test Bed (NSTB), which is an activity of the
Center for SCADA Security at Sandia. Sandia and Idaho National
laboratories were designated as co-leads of this effort. Other partners
include Argonne National Laboratory, Pacific Northwest National
Laboratory, and the National Institute of Standards and Technology. The
goals of the NSTB are to raise awareness of, and demonstrate the need
for, improved security. The approach is to demonstrate credible threats
against critical infrastructures and conduct vulnerability assessments
of SCADA systems. We also develop, in collaboration with industry, risk
mitigation strategies for current SCADA systems. We are developing new
architectures for future secure infrastructures, and we are supporting
the development of national guidelines and standards for secure SCADA
design and implementation.
Internal Sandia Programs
A number of Sandia facilities support the SCADA security effort,
including the Distributed Energy Technology Laboratory, which provides
a platform to test the control of operational generation and load
systems. We also have a Network Visualization Laboratory that provides
both visualization and network modeling capabilities, a Cryptographic
Research Facility that supports research and development of
cryptographic methods for SCADA networks, an Attack Resource Center
that provides tools to attack and analyze SCADA vulnerabilities, and an
Advanced Information Systems Laboratory that supports research and
development of intelligent agent technologies that may provide self-
healing infrastructures in the future.
Sandia also sponsors a nationally recognized College Cyber Defender
program that trains university students to protect electronic
information and defend computer systems and networks from cyber
attacks. The program encourages a pipeline of qualified candidates in
the fields of cyber security and protection to address Homeland
Security and national security needs.
Research
The Department of Homeland Security has funded the Institute for
Information Infrastructure Protection (I3P) to conduct research in
SCADA security in order to improve the robustness of the nation's
interdependent critical infrastructures. Sandia is the team lead for
this project, which includes faculty and staff from ten institutions
individually recognized for their expertise in cyber security and
critical infrastructure research: Sandia, University of Virginia, New
York University, University of Tulsa, Pacific Northwest National
Laboratory, Massachusetts Institute of Technology's Lincoln Laboratory,
SRI International, MITRE, University of Illinois at Urbana-Champaign,
and Dartmouth College. The institute is presently researching the
following six high-priority tasks:
Task 1: Assess dependence of critical infrastructures on SCADA
and its security.
Task 2: Account for the type and magnitude of SCADA
interdependencies.
Task 3: Develop metrics for the assessment and management of
SCADA security.
Task 4: Develop inherently secure SCADA systems requirements.
Task 5: Develop cross-domain solutions for information sharing.
Task 6: Transfer technology of these solutions into industry.
The institute represents the type of collaboration needed among
private stakeholders, academia, government agencies, and national
laboratories to solve the complex problem of SCADA security.
Suggestions for Addressing Critical Infrastructure Control System
Problems
Private industry owns about eighty-five percent of U.S. critical
infrastructure assets. Industry, therefore, has a key role in
implementing protection strategies. Currently, the business case (i.e.,
return on investment) for industry to invest in increasing the security
of their information systems has not been convincingly made. Part of
the reason is that no one has been able to clearly define a specific
threat. In the past, industry has demonstrated its willingness to
invest in protection when faced with a specific threat. The best
example of this is the hard work and dedicated effort that industry
provided to counter the Y2K threat.
Although we know that many threats exist, specific details are
elusive. It may be that we will need to take a consequence-based
approach--rather than a threat-based approach--to provide the rationale
for the business case. This approach would involve identification of
specific portions of information systems affected by specific attacks.
It would require vulnerability assessments, analyzing the consequences
of disruptions in economic terms, and defining and implementing
optimized protection strategies based on risk assessments. The national
laboratories use sophisticated means to develop simplified assessment
and risk survey processes, like the RAM-W work at Sandia. Risk
assessment methodologies can quickly and more broadly identify the
current security conditions and help decision-makers plan the most cost
effective steps to improve a particular infrastructure's security
posture. Increased emphasis should be placed on public-private
partnerships in order to make this process efficient.
When considering solutions, the difference between levels of
threats needs to be considered. The current emphasis by industry is to
try to eliminate inherent vulnerabilities that are present in all
networked computer systems. Hackers and hacker coalitions view these
vulnerabilities as low-hanging fruit. They exploit them to steal
information and identities and/or to deny or disable processes. There
is recent evidence that organized crime is also exploiting these
vulnerabilities for extortion purposes. Academia and the industrial
information security groups are working to provide technology solutions
to counter the lower level threat. Until those solutions arrive, all
critical infrastructure providers should apply best practices for
defense against inherent system vulnerabilities. These practices should
include development of security policy as well as technology solutions
to provide a sustainable security environment.
At the same time, terrorists and nation states are developing
attack methods that are much more sophisticated, often covert. We need
new efforts to identify, characterize, and counter these threats.
Perhaps this is the proper role for government agencies with technical
support from the national laboratories. In that case, the government
agencies and national laboratories that are working on high-end
defensive solutions will need to establish a plan for technology
transfer to industry, because the methods used by today's sophisticated
adversary will at some point be available to the lower level threat
community.
It is clear that successful defense of the nation's infrastructure
will require increased interagency cooperation. For example, the
Department of Defense (DoD) has a vital interest in the reliable and
secure operation of the nation's critical infrastructures because the
U.S. military depends on both domestic and international
infrastructures to conduct its missions. Thus the DoD has a keen
interest in protecting the SCADA systems that monitor infrastructures,
and cooperation with other U.S. agencies will be vital to its mission
success.
The Department of Homeland Security (DHS) is already working with
the DOE on cooperative interagency projects like the National SCADA
Test Bed and the DHS's SCADA security programs. These two agencies
should continue their cooperative efforts to ensure that work is
coordinated effectively, all threats are considered, the best
technology is used, and duplication of effort is avoided. The
collaborations and partnerships called for in Homeland Security
Presidential Directive 7 (Critical Infrastructure Identification,
Prioritization, and Protection), along with the roles and
responsibilities described there, are key to accomplishing these goals.
Recommendations
Reaffirm the concept of public-private partnerships
and encourage participants to share information on threats,
vulnerabilities, consequences of outages, training, and
technology. Extend these partnerships to assist industry in
making the business case for investments in security upgrades.
Increase funding for improvements in cyber security
technology, for example: tools for high speed intrusion
detection systems, software assurance, attack attribution and
trace-back, security modeling of existing and proposed SCADA
systems, network visualization for mapping cyber disruptions,
triage of threat scenarios across many vectors, and methods for
assuring the reliable performance of COTS products.
Establish and fully fund additional work that provides
defense against sophisticated threats.
Continue Congressional support of the initiatives and
directives described in the National Strategy for the Physical
Protections of Critical Infrastructures and Key Assets, the
National Strategy to Secure Cyberspace, Homeland Security
Presidential Directive 7, the Interim National Infrastructure
Protection Plan, and associated Sector Specific Plans.
Thank you, Mr. Chairman. I would be pleased to respond to any
questions you may have.
ATTACHMENTS
Supplemental Statement of Dr. Samuel Glenn Varnado
Sandia National Laboratories
Summary of Major Points
The nation's infrastructure is highly vulnerable to
cyber threats. Supervisory Control and Data Acquisition (SCADA)
systems are prime targets for hackers, terrorists, and nation
states.
U.S. computer networks are under daily attack.
Adversaries are becoming more sophisticated. We are seeing
structured, well-resourced attacks that are designed to steal
information or disrupt and/or deny processes.
Information technology vendors release four new
vulnerability announcements each day. At the same time, new
attack methods are proliferating. For example, Super Slammer, a
fast worm, infected 60% of the Department of Defense's (DoD's)
NIPRNET (Unclassified but Sensitive Internet Protocol Router
Network) machines in eight minutes.
Most of the current emphasis in the cyber security
community is on responding to hacker incidents. This effort is
necessary and useful; however, the work has a short-term focus.
We must mature our thinking in the area of enterprise-wide
network defense strategies. In addition, more complicated
threats such as terrorism and nation state actors must be
addressed.
We have no alternative to the use of Commercial Off
the Shelf (COTS) products in all our information systems. Most
of these hardware and software products are manufactured in
countries whose interests do not always align with those of the
United States.
We must understand that we will be attacked. What are
the implications of that understanding, and what strategies do
we have in place to operate through the attacks in order to
implement recovery and response activities?
We need to expand our investment in cyber security
technology development in order to address the new threat and
vulnerability environments.
We must encourage more public-private partnerships to
share threat, consequence, and vulnerability data and to
implement cost effective security solutions.
We must help industries develop a business case for
their investment in SCADA security.
Sandia National Laboratories has been working to
improve the security of SCADA systems for over ten years. We
have invested laboratory directed research and development
(LDRD) and other appropriate sponsor-provided funds into
technologies that have direct application to homeland security
and infrastructure protection.
Mr. Lungren. The Chair would now recognize Dr. K.P. Ananth,
Associate Laboratory Director for National Homeland Security at
the Idaho National Laboratory, to testify.
STATEMENT OF K.P. ANANTH, ASSOCIATE LABORATORY DIRECTOR,
NATIONAL AND HOMELAND SECURITY, IDAHO NATIONAL LABORATORY
Mr. Ananth. Thank you, Chairman Lungren and distinguished
members of the homeland security subcommittee. I am K.P.
Ananth, Associate Lab Director for National and Homeland
Security at the Idaho National Laboratory, a DOE national lab.
It is a pleasure for me to appear before you to represent the
work carried out at INL in support of our national efforts to
protect critical infrastructure. In this testimony I will give
you a short summary of our unique capabilities related to
SCADA, critical infrastructure protection, and cybersecurity,
the work we do and the challenges we face.
For the last half century, INL has played a key role in the
energy security and national security of the U.S. through its
pioneering work in nuclear reactors, nuclear power and nuclear
ship propulsion, and, as a result, developed a significant
infrastructure with one-of-a-kind test beds and facilities on a
secure 890-square-mile complex in Idaho. The written testimony
provides details on many of the facilities, but I will focus
here on those assets directly related to improving
cybersecurity and a critical infrastructure protection mission.
Process control systems in SCADA at the INL include a 61-
mile, 138-kilovolt transmission line with seven substations and
a power distribution control center, a pilot chemical plant,
and significant cybersecurity capabilities. We have 10 SCADA
test beds with plug-and-play capabilities that a system might
need for evaluation. These test beds are secure to protect
vendor systems and information and have connectivity to the
test range.
Additionally, we work with the global commercial vendors
such as ABB, AREVA, GE, Siemens and others, and we enable our
work through these vendor systems to look at the system
vulnerability and to improve cybersecurity.
Additionally, INL's low radio frequency background,
combined with our NTIA status and access to major telecom
vendors, enables INL to address risks and improve robustness of
communication links. This portfolio of unique test beds
complemented with our experienced staff and our collaborators
in the national laboratories, academia and industry serve as a
national resource for critical infrastructure protection.
Now I will touch upon the key programs we have and results.
The DHS program known as US-CERT Control Systems Security
Center is aimed at improving control systems security across
all critical infrastructure sectors. Key accomplishments
include design of a cybersecurity framework and self-assessment
tool for industry that is being validated by industry and NIST.
This will be piloted in fiscal year 2006.
We support US-CERT in handling control systems-specific
incidents and events, preparing bulletins and support for
reported events. We have expanded the cyber test bed with three
fully functioning systems and tested control systems of vendors
showing vulnerabilities and shared them with industry. We have
provided training and tabletop demonstrations at 9 U.S.
locations to 460 end users.
The DOE program known as the SCADA Test Bed performs
testing and analysis focused on the energy sector. We have
identified key vulnerabilities in four major control systems
used in the electric sector and worked with vendors to develop
fixes. We have shared findings with over 200 representatives of
100 major industry owner user groups through invited
participation. We provided SCADA security NERC-certified
training and other courses to over 350 participants. Through
these programs we have helped industry develop and deploy more
secure digital control SCADA systems and evaluated technology
from providers representing 80 percent of the control systems
market for the electric grid.
Now I will move on quickly to the challenges. Increased
connectivity. As my colleague mentioned here, control systems
today are susceptible to security threats due to open industry
protocols and access to control systems information via public
networks, legacy systems. Many of the older control systems
with long life cycles did not consider cybersecurity; hence,
they are vulnerable.
Deregulation. Utility deregulation has increased the number
of entities involved in the power life cycle, from generation
to transmission, distribution, marketing and billing.
Consequently there is increased connectivity and increased
potential for cyberattacks via corporate networks.
Offshore reliance. Again, cost pressures and technology
support constraints have driven companies to go abroad, again
causing security vulnerabilities.
And the need for information sharing is also critical.
Although these challenges are numerous, they are
surmountable, and we have got some recommendations that are in
the written testimony that you will see.
Mr. Chairman and distinguished members of the group, we
invite you to visit Idaho, see the test bed and the work we do
in supporting the Nation's infrastructure problems. Thank you.
Mr. Lungren. Thank you very much, Doctor.
[The statement of Mr. Ananth follows:]
Prepared Statement of Dr. K.P. Ananth
Chairman Lungren and distinguished members of the Homeland Security
Subcommittee:
I am Dr. K. P. Ananth, Associate Laboratory Director for National
and Homeland Security at the Idaho National Laboratory (INL), a DOE
national laboratory. It is a privilege and honor for me to appear
before you to represent the work being carried out at INL in support of
our national efforts, undertaken in both the federal and private
sectors, to protect U.S. critical infrastructure. In this testimony, I
will give you a brief background on INL and its mission, and a summary
of our unique capabilities as they relate to Supervisory Control and
Data Acquisition (SCADA), Critical Infrastructure Protection (CIP) and
Cyber Security. I will also discuss key federal and commercial programs
carried out at the Laboratory to support industry and end users, and
identify the challenges we face along with some recommendations.
INL and its Mission
The Idaho National Laboratory had its origin as the National
Reactor Testing Station in 1949 in Idaho Falls with a mission to
design, engineer, develop a prototype, and test an electricity
producing nuclear reactor. Within two (2) years, in December 1951, INL
successfully demonstrated the first electric power reactor and, soon
thereafter, developed the first prototype nuclear reactor for the
nuclear submarine Nautilus. For more than 50 years, the laboratory has
been a critical asset within the National Laboratory system as an
engineering, prototyping and testing resource, with 52 reactors built
and operated on the 890 square mile reservation in southeastern Idaho.
Beginning in the 1950s, the Laboratory began to support major
Department of Defense programs, including training of thousands of Navy
nuclear operators; earlier the Laboratory was involved in the
development and testing of naval guns and ordnance. In 1985, the
Laboratory was selected to produce armor for the Army's Abrams tank
using depleted uranium, and earlier this year we successfully completed
our twentieth anniversary on the program.
To support these varied missions, INL has developed a significant
infrastructure on the Idaho desert. INL carries the distinction of a
vast, remote, and secure heavily-invested site complex with ``one-of-a-
kind'' test beds and facilities for nuclear research and development
(R&D), explosives detection and testing, unmanned aerial and ground
vehicles payload testing, physical security, cyber security and
critical infrastructure protection. Mindful of the rich assets at INL,
the Department of Energy issued a Request for Proposal (RFP) in 2004 to
manage and operate INL with the mission of ensuring the nation's energy
security with safe, competitive, and sustainable energy systems and
providing unique national and homeland security capabilities. Two areas
were specifically called out within national and homeland security for
the Laboratory: nuclear nonproliferation and critical infrastructure
protection. On February 1, 2005, the new contract to operate the
Laboratory was implemented, making the critical infrastructure
protection mission of the Idaho National Laboratory unique within the
National Laboratory system. We are hard at work fulfilling this
mandate.
Today I will focus on how we are leveraging our efforts with DHS
and DOE in the area of improving control systems security across all
critical infrastructure sectors by reducing cyber security
vulnerabilities and risk.
INL's Unique Assets
With more than five decades of experience in establishing,
developing and maintaining critical infrastructure systems, INL has
created several recognized and integrated capabilities to provide real
solutions to our customers in critical infrastructure protection and
cyber security. INL has focused in three major areas--process control
systems, cyber security, and wireless technology.
Process Control Systems (PCS) and SCADA--Our location and
operational infrastructure provides the ultimate proving ground for
analysis and assessment of real-world critical infrastructure
components. INL has become the logical home for significant portions of
the National SCADA Test Bed and has become the focal point for research
and testing of control systems and cyber security with a direct benefit
of increasing the security of these systems. INL operates a power
distribution control center, a pilot chemical plant, and 61 miles of
138 kV transmission line with seven substations and a dedicated control
room on our 890 square mile site. It is the combination of this
infrastructure, a program with current access to commercial control
systems from principal global vendors (e.g., ABB, AREVA, GE, METSCO,
Micro Motion, [Emerson], Rockwell Automation, Siemens), and our
research expertise and partners that enables us to conduct offline and
full-scale testing in a real life environment. This unique capability
is helping to research and develop solutions that will strengthen our
nation's industrial control systems and physical components of our
infrastructures from attacks by viruses, hackers, and terrorists.
Cyber Security--the INL Cyber Security Group's intimate familiarity
with various hacker methodologies enables us to generate exploits and
assessment tools for use in testing the security of Critical
Infrastructure control system environments. Focused on multi-tier
attack vectors and full spectrum threat actors, the team provides a
credible representation of cyber threats and then conducts cutting edge
research into advanced mitigation strategies and solutions. Coupled
with our academic and industry partners in this area, we are striving
to effectively address current challenges while advancing the state-of-
the-art in detecting hacker signatures. We have invested resources to
explore the cyber security vulnerabilities of Portable Electronic
Devices (PEDs) technology. INL is pursuing commercial and government
partnerships to address vulnerabilities in PEDs technology because
these devices are becoming more prolific and have crept into new
control systems.
Wireless Technology--INL's Wireless Test Bed and telecommunications
infrastructure provides access to advanced, next generation
communication technology and current communication systems to analyze
vulnerabilities, analyze new protocols and operational performance, and
develop risk mitigating solutions. INL's location providing a low RF
background, our National Telecommunications and Administration (NTIA)
experimental radio station status, full-scale isolated communications
networks, and ability to connect to functional systems has attracted
industry (e.g., Bechtel Telecommunications, Nokia, AT&T Wireless) and
government customers. Bechtel Telecom, through a Cooperative Research
and Development Agreement (CRADA), has made a significant investment at
the Laboratory in this area. These attributes afford us the unique
opportunity to holistically analyze both performance and risk of entire
systems, develop wireless security solutions for our nation's complex,
interconnected infrastructures, and improve robustness of communication
links for emergency responders.
The importance of these core assets can not be overlooked,
representing a national resource that provides access to control system
hardware and applications, functioning transmission and distribution
assets, wireless local and metro area networks, advanced radio,
microwave, fiber optic and satellite communications, mesh networks and
personal electronic devices (PEDs). Additional assets include unmanned
aerial vehicles (UAVs), explosives detection, testing and blast
mitigation systems. Perhaps more importantly, our current network of
industry participants and top shelf researchers across the nation
enable INL to address the most challenging issues in CIP.
These are the elements--housed in our comprehensive test range,
designed to be full-scale in nature, representative of real world
infrastructures and capable of being isolated--that uniquely position
the federal government, national laboratories, and industry to be
successful in identifying and managing risk to our nation's critical
infrastructure. To the best of our knowledge, there is no similar
facility in the world. And, the cache of over 100 experienced
scientists, engineers, and technicians working in INL's SCADA/Cyber
Security groups are aware of the great responsibility that comes with
managing these resources and the significance of our mission to assist
in securing the control systems of our nation's critical
infrastructure. With this knowledge, we have focused on developing
extensive collaborations on our programs and continually strive to
bring the best-in-class institutions to help in developing solutions to
this complex challenge. Our collaborators in this area include other
national laboratories, National Institute of Standards and Technology
(NIST), American Society of Mechanical Engineers (ASME),
Instrumentation Systems and Automation Society (ISA), Carnegie Mellon
University (CMU), Dartmouth University (DU), University of Idaho (UoI),
British Columbia Institute of Technology (BCIT), and others such as
North American Electric Reliability Council (NERC), Electric Power
Research Institute (EPRI), Chemical Industry Data Exchange (CIDX),
Decision Analytics Corporation (DAC), KEMA Consulting and Bearing
Point.
Key Programs Conducted at INL and Results Achieved
Our two primary programs in Cyber Security and Critical
Infrastructure Protection are with the Department of Homeland Security
National Cyber Security Division and Department of Energy Office of
Electricity Delivery and Energy Reliability. INL is supporting both
programs with a team of talented people from other national labs,
academia and industry based on their best-in-class core competencies
and the needs of the program.
The DHS program is known as the ``US-CERT Control Systems Security
Center (CSSC) Program.'' This program is aimed at improving control
systems security across all critical infrastructure sectors by reducing
cyber security vulnerabilities and risk. One of the key tasks of this
program was the design of a cyber security protection framework
consisting of a comprehensive set of requirements, graded
recommendations/solutions, and automated self-assessment tools for all
sectors to use to enhance the security of their control systems (e.g.,
SCADA, DCS) against cyber attack. The draft framework was issued in
July 2005 and reviewed with 20 industry control systems and cyber
experts; and a second review occurred in August with several key
industry security managers. Comments to date have been:
``. . . .framework provides a centralized, organized approach
to Control System security. . .''
``. . . .provides actionable recommendations. . .''
``. . . .provides a benchmark and metrics for cyber security
protection. . .
``. . . .will help consolidate the efforts by the Standard
bodies
. . .''
``. . .provides for cross platform standardization across
vendor products
. . .''
``. . . .impressed with the automated self-assessment tools
that will measure
improvement
over time
. . .''
We have plans to work with NIST and ISA over the next three months
to assist us in implementing the cyber security framework for self
assessment. We will also work with facilities in several key sectors in
FY-06 to pilot and validate the framework. A key component of the self
assessment will be a risk reduction tool that helps companies
prioritize vulnerabilities that are found when assessing requirements
and potential consequences.
Additionally, the program also developed a quick response cell to
support US CERT in handling control system specific incidents/events.
We have assisted in preparing cyber security bulletins and providing
Tier II support for reported events to the US-CERT.
Over the last two years, we have collaborated with DHS and DOE to
significantly increase the capabilities of our extensive cyber test
bed. This capability includes ten (10) SCADA test beds and three (3)
fully functioning systems that are ready and are currently testing
vendor systems and specific tools to reduce cyber vulnerabilities. On
the CSSC program, we are currently testing three (3) vendor control
systems and have already identified significant vulnerabilities on the
first two systems. The vendors are evaluating the results and our
recommendations.
The purpose of this program is to reduce risk to key infrastructure
from cyber attack by enhancing the security of control systems. To that
end, we have developed a risk assessment methodology for control
systems to measure vulnerability reduction and we have developed
decision analysis tools. We have started validating these tools by
analyzing test results and attack scenarios.
Our industry outreach efforts provide unique training by
demonstrating how an attack may propagate through the business system
to critical control systems with an emphasis on how to mitigate the
effects of such an attack. These awareness demonstrations and training
activities are ongoing with positive feedback from industry and
government participants. The tabletop demonstrations have included live
demonstrations of attacks/effects on small scale representative control
systems for chemical and electric system processes and demonstrations
of attack mitigation strategies. We have held these demonstrations at
nine (9) venues across the U.S. with over 460 end users participating
from a wide variety of industries to include control systems/cyber
security organizations and federal, state and local government
agencies.
Through this program, we are also providing SCADA and process
control security training for the protection of dams and hydroelectric
facilities to system users in the Department of Interior's Bureau of
Reclamation.
The DOE program, known as the ``National SCADA Test Bed (NSTB)
performs testing and analysis of SCADA systems representative of those
used throughout the energy sector to identify, validate and reduce
cyber vulnerabilities. The second objective is to identify best
practices for design and deployment of secure control systems and to
support institutionalization of those best practices in government and
industry standards. The NSTB is a joint effort between Sandia National
Laboratory and Idaho National Laboratory. The NSTB effort is managed by
the INL and includes, Pacific Northwest National Laboratory (PNNL),
Argonne National Laboratory (ANL), and the SCADA vendor community (ABB
Network Management, AREVA T&D Automation, GE Energy Management Systems,
Siemens Power Transmission and Distribution), as well as computer
system vendors such as IBM, HP, and Sun Systems. Key accomplishments on
this program include:
The NSTB has identified SCADA vulnerabilities in the
four systems INL has tested, worked with the SCADA vendors to
define/develop fixes where needed, and verified the fixes
through follow-on testing. SCADA vendors have improved new
releases and developed patches to mitigate significant security
weakness. These risk reducing actions will directly benefit
many of the nation's critical infrastructure organizations.
We have shared the findings from these SCADA system
vulnerability assessments, in various levels of detail, with
over 230 representatives from 100 major industry owner/user
organizations through invited presentations at SCADA vendor
users' group meetings.
We have issued detailed test reports of the SCADA
assessments to the respective vendors. One of the vendors is
sharing their assessment report, under tight non-disclosure
agreements, with all interested users.
Through the participation of SCADA vendors who have
been willing to loan their systems to INL on the NSTB program
for an extended time, we have established an extensive,
representative environment for searching out typical security
vulnerabilities and for testing solutions.
We developed and presented a NERC-certified training course on
SCADA security. Based on feedback from the initial presentation of
various courses (NERC and others) to over 350 participants, we are
expanding the content and are now responding to requests for additional
presentations.
Commercial Programs--INL has helped industry develop and deploy
more secure digital control/SCADA systems, through vulnerability
discovery, validation and mitigation, standards development and secure
software technology.
Specifically, the INL managed National SCADA Test Bed Program
(NSTB) has worked with global control system software vendors to
promote more secure , innovative installation and implementation of
their products, where such efforts are consistent with recognized
industry guidelines and best practices. The program has discovered
existing weaknesses in deployed systems as well as design weaknesses in
future control systems. The program has evaluated technology from
providers representing 80% of the electrical grid control system
market, working closely with engineering teams of four (4) global
providers.
We have worked with control system owners and operators across
multiple sectors to evaluate and enhance security of existing
technology deployments. These companies took advantage of the unique
knowledge-base and trusted relationships at the Lab as an important
element to their overall approach to critical systems risk management.
Companies have also turned to us when things go wrong with the systems
to assist in evaluating particular events to determine if directed or
non-directed attacks might have occurred.
With most of the critical infrastructure residing in the private
sector we felt it was appropriate to submit just a few comments from
the asset owners themselves. These perspectives come from private
sector organizations from the trenches to the executive offices best
demonstrating the value of government sponsored CIP initiatives at INL:
1. David Norton, Transmission IT Security program manager for
Entergy--New Orleans (the second largest generator of
electricity in the U.S. delivering electricity to 2.7 million
customers), wrote ``We are in dire need of INL, its mission,
and its uniquely qualified staff. I know of no other entity in
North America doing anything like what they are doing in the
field of SCADA control system security, and certainly not to
the level of excellence that I and my peers in the industry
have witnessed.''
2. Cheryl Santor, Information Security Manager, Metropolitan
Water in California (one of the largest water systems servicing
5,200 square miles in Los Angeles, Orange, San Diego,
Riverside, San Bernardino and Ventura Counties with 18 million
customers), wrote ``The INL provides a knowledge base from
which all organizations using SCADA and Process Controls can
benefit. . . .in order to secure their critical resources.''
3. Phil Harris, CEO of PJM (Ensuring the reliability of the
largest centrally dispatched Control area in North America by
coordinating the movement of electricity in all parts of
Delaware, Illinois, Indiana, Kentucky, Maryland, Michigan, New
Jersey, North Carolina, Ohio, Pennsylvania, Tennessee,
Virginia, West Virginia, and the District of Columbia), wrote
``PJM feels it is important that the Electric Sector, as a
Critical Infrastructure support INL and the work they do. There
is no substitute or other entity that is providing such quality
service of such national importance.''
4. Another utility security executive from American Electric
Power recently testified to the value provided by INL through
the DHS and DOE program: ``The electricity industry is
interested in continuing to work closely with DOE on the work
being done at the Idaho National Laboratory. We believe it
holds great promise as one of the best and most efficient means
of stimulating research and developing technical solutions to
the present shortfalls in cyber security.'' [Hearing Before the
United States House of Representatives Science Committee,
September 15, 2005].
Key Challenges in CIP and Cyber Security
As a result of operating and testing infrastructure systems,
working with control system vendors and end users, INL is keenly aware
of the key challenges in protecting critical control systems and the
potential solutions to these complex challenges to ensure the security
of our nation's critical infrastructure.
Increased Connectivity--The use of open systems and more
common technology combined with greater system access and available
system knowledge has changed the risk profile of SCADA systems. These
systems evolved in a less connected world relying on proprietary
technologies which provided a sense of ``security through obscurity''
in the past. The control systems of today are more susceptible to
security threats than before with SCADA vendors increasingly moving
toward open industry standard protocols and platforms, system owners
and operators providing greater access to market and accounting
systems, regulatory requirements to share information and make systems
available to all market participants and the greater use of public
networks and wireless communications.
Interdependencies--A further challenge arises from the
reliance on telecommunication as an integral part of the overall
control system. If SCADA and Energy Management Systems (EMS) are the
brain stem and receptors of a control system, then Telecommunications
represents the intricate network of nerve pathways that connects these
operational assets, providing the means by which to deliver the control
instructions and update system status. [The following provides a useful
reference: Cyber Security: A Crisis of Prioritization, President's
Information Technology Advisory Committee, February 2005]
Complexity--A particular challenge is the complex and
interconnected nature of critical control systems which can be found
across many of the critical infrastructure sectors from directing
advanced manufacturing systems to controlling the North American
electric grid. If we focus on energy production and delivery, we find
Process Control Systems (PCS) and specifically SCADA systems are used
extensively throughout the electric, oil, and gas sectors to monitor
and control processes that generate, transmit, transport and distribute
energy.
Legacy Systems--A significant portion of the control system
technology in place today in many installations is old. These legacy
systems were designed to operate over long lifecycles and were not
designed with cyber security in mind. Hence, they are vulnerable to
cyber attack and, in many cases, difficult to protect. In order to
significantly lower the risk, we need to understand legacy system
vulnerabilities and develop cost effective means to mitigate them
without relying on new system deployments.
Deregulation--Market forces, to include deregulation in the
electric utility industry have increased the number of entities
involved in the power life cycle from generation through transmission,
distribution, metering, and billing; thus increasing reliance on and
accuracy of information from third parties. Correspondingly, this has
come with increased connectivity with outside vendors, customers, and
business partners which have eroded the sanctity of the network
perimeter. More connections through the perimeter inherently introduce
more threats into the corporate networks.
System Accessibility--The convergence of power company
networks and the demand for remote access to these systems has rendered
many SCADA systems accessible through non-SCADA networks. Specifically,
connections between the grid and corporate networks for reporting
purposes and outage management interfaces have the potential to expose
the grid network to the threats experienced by the more common business
network. [The following provides a useful reference: U.S.-Canada Power
System Outage Task Force, August 14th Blackout: Causes and
Recommendations, April 2004].
Offshore Reliance--Cost pressures and technology support
constraints have increased reliance on offshore development and system
maintenance, thereby increasing the risk of intentional or
unintentional security vulnerabilities. This risk is amplified as a
result of ineffective/non-enforceable cyber laws in the respective
offshore countries.
Information Sharing--Finally, competitive pressure, legal
liability risk and the lack of information protection mechanisms pose a
significant barrier to information sharing between critical
infrastructure stakeholders. This has significantly impeded the
discovery and understanding of control system vulnerabilities, as well
as the reporting of real-world incidents. [The following provides a
useful reference: CRS Report for Congress – Government Activities
to Protect the Electric Grid, October 2004]. On the other hand, the
knowledge revolution that has accompanied the Internet makes it easy to
locate specific information regarding SCADA and automation systems. For
example, ``over 90% of major SCADA and Automation vendors have all of
their manuals and specifications available online to the general
public'' (SCADA Security Strategy, PlantData). Easy access of such
information to potential threat actors is a concern.
Recommendations
These challenges, although numerous and complex, are surmountable.
There is an urgent need to accelerate the research, development,
testing, and application of advanced control systems to enhance cyber
security across the energy and other sectors. This need transcends
individual companies, energy subsectors, and even the private sector.
Toward this end, the Department of Homeland Security and the Department
of Energy are supporting programs to facilitate and support risk
reducing solutions. We, at INL, are focused on providing solutions to
this key national need and have some recommendations for meeting the
challenge.
SCADA/Cyber/Telecom Interconnect--We, as a nation, should develop an
interdependent and inclusive view of control systems to include not
only the SCADA systems but the cyber and Telecommunications functions
that support them to ensure secure electrical power and industrial
processes. SCADA, Cyber Security, and Telecommunications are areas
where we must integrate research and testing efforts to understand how
vulnerabilities impact the entire system. We at INL are already engaged
with the telecommunication firms on interoperability and bandwidth
issues, and we see the SCADA/Cyber/Telecom interconnectivity as the
next area of pursuit.
The 21st Century could be characterized as a globally
interconnected ``flat world'' (courtesy of Tom Friedman), which means
hierarchical systems have to yield to horizontal and partnership-based
enterprises. To that end, critical infrastructure protection, cyber
security, and telecommunications particularly call to attention the
interdependence between providers and markets so industries have a
responsibility to work across sectors, and the same holds for the
federal government. Furthermore, in the event of a manmade or natural
disaster as in Katrina, active coordination across sectors is vital for
timely response and expeditious recovery.
Minimum Standards--The electric sector, being at the hub of all, is
active in securing its cyber and physical resources. Interim cyber
security standards are in place in the electric sector, and they are
moving through the approval process for a permanent, more expansive CIP
standard. The final product should strengthen cyber security across the
electric sector and lay the groundwork for greater collaboration
between industry and government. Similar efforts are underway through
CIDX and much work remains to be done in all sectors of our
infrastructure.
Develop Risk Assessment Tools--The federal government should continue
to invest in the development of tools and provide required information
to assist control systems security professionals to identify and
address risk. Education and awareness efforts should be focused on
developing an accurate understanding of risk to control systems. The
NSTB Program and the CSSC program are both actively addressing this
need and risk mitigation steps are beginning to be implemented at the
user level.
Fixing Legacy Systems--Some type of incentive, either at the vendor
level or user level, will go a long way to implement cyber security in
legacy process control systems. Coupled with independent third party
testing of the control system, through programs such as NSTB and CSSC,
legacy systems could be upgraded with protective measures.
Information Protection--The electric infrastructure is one of the most
critical infrastructures servicing the nation and maintaining our way
of life. Certain technical, architectural and operational aspects and
details must be kept secure so they will not be inadvertently disclosed
to those who would try to disrupt or destroy our social, political or
economic fabric. Yet there is a need to share the security aspects of
the information with government and industry peers for benchmarking
purposes while preserving competitive advantages. The same challenge
applies to other sectors as well. This is an area where the use of
trusted independent third party entities might prove beneficial and
acceptable to all parties and merits further discussion.
Concluding Statement
Mr. Chairman and distinguished Members of the Committee, we at
Idaho National Laboratory are fully committed to deliver on this
important national mission, and along side DHS, DOE, and industry, we
will strive to make our Laboratory the Center of Excellence in critical
infrastructure protection to help end users. We welcome you to visit
the Idaho National Lab to see firsthand the solutions we are providing
to make our infrastructure safer. Again, I thank you for the
opportunity to share these comments with you.
Mr. Lungren. The Chair would now recognize Dr. William
Rush, institute physicist at the Gas Technology Institute, to
testify.
STATEMENT OF WILLIAM RUSH, INSTITUTE PHYSICIST, GAS TECHNOLOGY
INSTITUTE
Mr. Rush. Good afternoon, Mr. Chairman and members of the
committee. I would like to thank you very much for letting me
testify on what I think is a really important topic. I am Bill
Rush. I hold a Ph.D. in physics, and for the past 27 years I
have been with the Gas Technology Institute, or GTI. I also
chair the American Gas Association's Encryption Working Group,
which is charged with developing cryptographic protection for
SCADA communications.
Today I am going to update you on the nuts and bolts of
what it is that we have done to protect against cyberattack and
recommend some specific steps for improving SCADA security.
As you know, attacks against SCADA are of concern because
SCADA is the remote control, if you will, of a network. It
controls the circuit breakers and the valves. It is the actual
``reach out and grab things'' part of the system. Most systems
were designed before security was regarded as a serious concern
and as a result are poorly protected against cyberattack. One
team of U.S. network experts was into a SCADA system within 15
minutes.
Can cyberattack have real consequences? Absolutely. As Dr.
Varnado pointed out, a 3-kiloton explosion, to put that into
more usable or more familiar terms, that is about 1,000 times
as powerful as the explosion that blew up the Murrah Federal
Office Building in Oklahoma City.
SCADA information has been found on captured al-Qa'ida
computers. Three weeks after the 9/11 attacks, the American Gas
Association chartered the AGA 12 Working Group to develop a
standard to protect SCADA communications. The drawing that we
have up here indicates basically how it works. What you do is
originate a command, such as open the switch inside a secure
facility. It then gets sent into a cryptographic module which
changes the message, and as you can see across the bottom, it
can't be read by anybody without a special number that is
called a key. When it shows up on the other end, it is
decrypted by the same key and turns back into the message, open
the switch.
AGA 12 team is proud of its progress to date, but this is
not just a paper standard. This device that I have brought with
me, and you can see it afterwards, is an AGA 12-compliant
cryptographic module. This unit effectively slams the door in
the face of those who would attempt to penetrate the
communication networks of SCADA systems. Early versions of this
equipment has performed well in the field tests. This unit is
priced at about $500. It can be installed right now in most
SCADA systems that operate on low-speed links. Nationally labs
are in the process of evaluating its security level and its
performance. At least two manufacturers will market AGA 12
modules. No other standard groups can provide this protection
today.
While many groups contributed to AGA 12's success, none did
more so than the Navy's Technical Support Working Group, or
TSWG. TSWG funded GTI to work on AGA 12 full time. This allowed
us to move far faster than any other all-volunteer groups.
Note that AGA 12 is only one of dozens of groups who are
involved in developing standards. There is a significant risk
of developing conflicting standards. These volunteer groups
lack the resources to coordinate their efforts. The DHS Process
Control Security Forum and DOE's Roadmap are important examples
of government and private sector coordination in cyber
security.
Regrettably, AGA 12 has become a victim of its own success.
TSWG only funds prototypes until they succeed. When AGA 12
passed this milestone last May, both funding and progress
ceased with a serious loss of momentum. Our early success
obscured the fact that critical work remains. DOE is providing
some funding to go restart tests and to edit parts of AGA 12
for publication, but there is still critical work, including
developing a seal of approval conformance testing to show that
a product such as this really meets the standard, sort of a
Good Housekeeping seal of approval; next-generation designs to
work faster and at half the cost; a major pilot test to
validate that the technology really works; and remote key
changing so you don't have to send staff out when you make
changes; and forensic tools to find and prosecute attackers.
In summary, we make the following recommendations: Fund R&D
to develop protection against cyberattacks on the Nation's
critical infrastructure. Prevent loss of momentum by avoiding
program interruptions. This is very disruptive. Support the
coordination effort, such as the Process Control Security Forum
and the Roadmap. Complete the remaining AGA 12 work that I have
just outlined. Support other selected standards works in
addressing the many vulnerabilities that are beyond the scope
of AGA 12.
Mr. Chairman and subcommittee members, we applaud your
focus on securing our Nation's critical infrastructure,
especially the area of SCADA protection. I would be pleased to
answer questions afterwards. Thank you.
Mr. Lungren. Thank you very much, Dr. Rush.
[The statement of Mr. Rush follows:]
Prepared Statement of Dr. William F. Rush
INTRODUCTION
Good afternoon Mr. Chairman and members of the Subcommittee. Thank
you for the opportunity to address you today on this important topic.
My name is Bill Rush and I hold the position of Institute Physicist
with the Gas Technology Institute (GTI), where I have worked in the
field of natural gas technology research and development for 27 years.
GTI is a not-for-profit Research and Development institute
headquartered in Des Plaines, Illinois. I also am the Chairman of the
American Gas Association's SCADA Encryption Working Group. The American
gas industry has charged this group with developing cryptographic
protection for gas, water, and electric SCADA communications.
The focus of my testimony today is to update you on the steps the
American Gas Association AGA, GTI, and many other organizations have
begun to take to protect SCADA communications from cyber attack. At the
conclusion of my remarks, I will provide recommendations to the
Subcommittee on what actions can be taken to further advance the
security of industrial control systems for critical infrastructures.
SCADA SYSTEMS ARE OFTEN VULNERABLE TO CYBER ATTACK
Supervisory Control And Data Acquisition (SCADA) systems are an
important component of critical infrastructure. SCADA systems can be
thought of as the ``remote control'' part of most gas, water, electric,
and oil pipeline systems. SCADA Remote Terminal Units (RTUs) read the
pressures, voltages, temperatures, and flows at critical points
throughout the transmission and distribution portions of these critical
infrastructure networks and transmit this real-time data back to
central control rooms. They also operate valves, circuit breakers, and
switches and are thus critical equipment for control of the systems.
This remote control of unmanned facilities provides quick response to
changing situations, while providing cost-effective operations of a
multitude of critical equipment and stations, spread over a large
geographic area. Many SCADA RTUs have ``maintenance ports'' that enable
operators to change critical system parameters remotely, open or close
valves or breakers, or download new firmware. There are strong
similarities among gas, water, electric, sewage, and oil SCADA systems.
Process automation and control systems used in other critical
infrastructure applications, such as oil refineries and chemical
plants, may not have the long-distance aspects of SCADA, but share many
other characteristics.
The cost constraints under which SCADA systems operate determine
many of their security-related characteristics. Because SCADA systems
are expensive to replace, they have long life times--typically between
10 and 20 years. Consequently, many systems now in service have been
there for a long time and will remain as legacy systems for some time
to come. Consequently, today's SCADA systems are often based on
technology which is a decade old. In particular, many of these systems
operate at relatively low communication speeds over telephone modems,
speeds which most Internet users of today find unacceptably slow.
Because these systems were designed before critical infrastructure
security was a major concern, they often have significant
vulnerabilities to unauthorized electronic operations, referred to as
``cyber attacks''. Many of the systems do not have effective password
protection for access control or encryption for confidentiality of data
and commands. When they use dial-in telephone modems, they often can be
hacked from any computer with a phone modem. When the SCADA system uses
radio communication, the radio waves can often be detected and altered
by a third party with an appropriate, commercially available receiver/
transmitter. The question confronting skilled cyber attackers is less
``Can we enter the system?'' and more ``How long will it take us to
penetrate it?'' The North American Electric Reliability (NERC) is
concerned about the ability of an attacker to use the maintenance ports
to attack SCADA systems by making unauthorized changes in critical
system parameters. Information on American SCADA systems has been found
on captured al-Qa'ida computers.
Cyber attacks are not simply minor incidents involving mildly
annoying hackers, but can have significant operational, economic, and
safety consequences. A single example that underscores this point is
the Soviet Union's use of stolen American SCADA software during the
1980's. This code--which had been deliberately modified to cause harm
to a SCADA system--led to physical damage to the Soviet SCADA system
resulting in an explosion large enough to be photographed from space
and estimated at 3 kilotons TNT equivalent. (See ``At the Abyss: An
Insider's History of the Cold War'', Thomas C. Reed, Ballantine Books,
New York, 2004.) To put the 3 kiloton number into perspective, the
Murrah Federal Office Building bombing in Oklahoma City was estimated
at 0.002 kiloton and the Hiroshima nuclear bomb was between 14 and 20
kilotons. The salient point is that it clearly is possible to cause
significant physical damage to critical infrastructure if the SCADA
code can be modified.
AGA 12 IS A STANDARD TO PROTECT SCADA FROM CYBER ATTACK
Three weeks after the 9/11 attack, AGA chartered a working group to
develop a comprehensive standard that would use cryptography to protect
SCADA communications from cyber attack. This standard has been
designated ``AGA 12''. When it is completed, it will be a comprehensive
approach to SCADA cryptography. The charter instructed the working
group to develop a recommended practice for the gas industry and to
include water and electric SCADA systems as well. This approach also
applies to sewage and oil pipeline SCADA systems. This effort has made
such significant progress that we are now field testing commercial
prototypes of products that use cryptography to protect SCADA
communications.
As a standard, AGA 12 has several significant characteristics.
First, it is an open consensus standard that is designed to produce
interoperable cryptographic products. ``Open'' means that anyone can
use the standard to build equipment without needing to pay a royalty or
licensing fee. Open here also refers to the process by which anyone
with an interest in the topic can participate in developing the
document. The working group included this requirement to encourage
market competition to drive costs down, since no one has a monopoly
position. The open-source code for implementing AGA 12 is available for
free on the Internet. AGA 12 is a consensus standard because the
working group develops consensus among its members and the AGA
membership as well that its recommendations are indeed a sound
practice. Finally, the standard specifies a minimum level of
interoperability among products made by different manufacturers. Thus,
users will have a choice of suppliers. The standard also assures that
new products will remain compatible with earlier versions. Finally, AGA
12 provides strong protection; it is based on well-established NIST
encryption standards and has been examined for its ability to protect
against a wide variety of attacks.
AGA 12 is a suite of 4 documents, designated Parts 1 through 4. The
four documents address different aspects of SCADA communication
protection.
AGA 12, Part 1 (AGA 12-1) summarizes cyber security policies, the
background of the cyber security problem, and a procedure for testing
cryptographic protection systems. This document educates SCADA
operators on the need to do a risk assessment and recommends an
approach for those utilities whose risk assessment reveals a need to
protect their systems with cryptography.
AGA 12-2 is a detailed technical specification for building
interoperable cryptographic modules to protect SCADA communications for
low-speed legacy SCADA systems and dial-up maintenance ports.
AGA 12-3 will describe how to protect high speed communication
SCADA systems.
AGA 12-4 will describe how to build next generation SCADA systems
so that their cryptography will be compatible with the legacy systems;
this will ease the transition to the newer designs.
Parts 1 and 2 are close to completion. Parts 3 and 4 are in the
planning stage.
Figure 1 illustrates both the configuration of a SCADA system and
the scope of AGA 12. On the left is the Control Room, which is manned
around the clock and where critical operational decisions are made. On
the right is the ``Remote Terminal Unit'' (RTU), which is typically
unmanned and controls the sensors and actuators that operate the
critical infrastructure. Both the Control Room and the RTU are assumed
to be secure. The AGA 12 working group deals only with the issues of
security of messages while they are in transit over an insecure network
and leaves to others the responsibility for securing the rest of the
system.
It is important to recognize that while cryptographic protection of
SCADA communications is an important weapon in the arsenal of tools
that can protect SCADA, it is only one tool among many that are needed.
Cryptography can not provide any protection at all against many kinds
of attacks. In particular, it does not protect against jamming or
breaking the communication line, against physical attacks, or against
many kinds of insider attacks. Nor does it protect local facility
control systems1 that are often connected to SCADA systems, and usually
offer additional independent vulnerabilities to cyber attack. These
issues are being addressed by literally dozens of groups working in the
security area. While I am focused only on the AGA 12 effort, I am
pleased to report that there are so many security initiatives under way
that coordinating their work is a major challenge. I would call your
attention to both the Department of Energy's Roadmap to Secure Control
Systems in the Energy Sector and the Department of Homeland Security's
Process Control Systems Forum as good examples of how the Government is
working effectively with the private sector to advance and coordinate
the many security efforts that are now under way. I also call your
attention to the Instrumentation, Systems and Automation Society's
(ISA) ISA SP99 committee, ``Manufacturing and Control Systems
Security''. This is a broad industry wide automation and control
systems security standards effort that has published over 150 pages of
guidance on how to establish automation systems security programs and
available technologies to deal with unacceptable risks. Finally, the
National Institute of Standards and Technology (NIST) has produced many
standards on which AGA 12 has relied and operates the Process Control
Security Forum (NIST PCSRF) which continues to advance putting the
cause of cyber security on a firm basis.
AGA 12 SPECIFIES CRYPTOGRAPHY TO PROTECT SCADA COMMUNICATIONS
AGA 12 uses cryptography to protect SCADA communications. Figure 2
illustrates the basic idea of how this works. Data and commands (``Open
Switch'' in this figure) originate inside of a secure facility, as
illustrated in Figure 1. Prior to leaving the secure facility, the data
or command is sent to a ``SCADA Cryptographic Module'' (SCM) which
encrypts it. Essentially, this encryption step changes the message so
that it can no longer be read by anyone without a special number,
called a key. In operation, the encrypted message is sent over the
insecure network in an unintelligible form. When it arrives at the
designated secure facility, the key is used to decrypt the message,
returning it to its original meaning, ``Open Switch''.
The AGA 12 standard has gone to great length to assure that
encrypted messages are very difficult for potential attackers to use to
harm a system that uses SCADA. This ``link encryption'' approach has
been used successfully for many years by the financial community to
secure its transactions. While this discussion has only considered
making the message hard to read, AGA 12 also makes it difficult to
alter, forge, or record and replay a message. An important issue
associated with AGA 12 is how these secret keys are managed. The keys
must be changed periodically to prevent their being guessed or
compromised. Different keys are used for employees with different
responsibilities and different levels of authority. The authorization
to use keys must, for example, be changed if an employee leaves. It is
important to be able to do this without the expense of visiting the
many distant sites that may be controlled by the SCADA system.
Because of the long life of SCADA systems, the owners and operators
of these systems urged the working group to focus first on the
challenging problem of protecting legacy systems. Focusing on next-
generation SCADA systems first would leave the legacy systems
unprotected for many years. Protecting legacy systems, however,
required developing cryptographic modules that will support most of the
roughly 150 types of existing SCADA systems, each of which has a
different ``SCADA language'' and which operate at different
communication speeds and over a wide variety of communication media
(such as telephone, radio, and microwave.) The next steps are to
develop the same standard protection for high speed and next generation
SCADA systems.
AGA 12 HAS MADE RAPID PROGRESS FOR A STANDARD
AGA 12 has made rapid progress, given the constraints that an open
group is developing a consensus standard. This is a process that is
generally slow for two reasons. First, developing consensus among
users, manufacturers, and cryptographic experts on a difficult
technical task is a challenging task. Each group has different needs
and understanding levels for the standard. Second, most standards
development efforts are all volunteer activities. This limits the rate
of progress to what can be accomplished in an overload or spare time
mode by people with full-time job responsibilities.
Those of us who have participated in the AGA 12 process are proud
of the success we have achieved, for this is no longer just a paper
standard. AGA 12 Part 1 is in the final stage of balloting prior to
being adopted as an industry recommended practice. Two manufacturers
are offering or soon will offer cryptographic modules that comply with
AGA 12, Part 2. Early versions of this equipment have performed well in
field tests at actual gas companies. AGA 12 has entered the field test
stage at least 2 years ahead of any other group developing an open
standard for cryptographic hardware.
MANY GROUPS HAVE CONTRIBUTED TO THE SUCCESS OF AGA 12
Many groups have contributed to the success of AGA 12. No single
group did more to accelerate the work of AGA 12 than the Technical
Support Working Group (TSWG), a part of the Combating Terrorism
Technology Support Office. TSWG began support of cryptography for SCADA
systems with a project at GTI in 1998, well before terrorism was
recognized as a threat. While as previously mentioned, most standards
groups operate on an all volunteer basis, TSWG funded GTI to provide
full-time support by several people to work on AGA 12. This allowed us
to debate approaches, build models of the various ideas, test to see
what does and what does not work, write our results into the emerging
standard, and begin the cycle anew with a debate on the next issue.
In addition to TSWG support, several other government agencies have
contributed to the progress of AGA 12. The National Institute of
Standards and Technology provided funding to help develop a standard
test methodology for evaluating how much cryptography slows
communications in network. Sandia National Laboratories evaluated the
security level of the first version, work which led to several
significant improvements to AGA 12. Pacific Northwest National
Laboratory conducted a preliminary test on the impact of AGA 12 on
communication speed. Under DOE sponsorship, both of these laboratories
continue to do work on the security and performance of the AGA 12-
compliant cryptographic modules. These National Laboratory tests are
particularly important to the private sector's acceptance of the AGA 12
standard as both secure and functional.
In addition to government support, industry groups have helped.
Both AGA and the American Water Works Association Research Foundation
(AWWARF) have provided funding and substantial in-kind support for the
AGA 12 standard. GTI and the Gas Research Institute have funded the AGA
12 work as well.
Many private companies also supported the AGA 12 project. These
include Cisco, OPUS Publishing, SafeNet Mykotronx, TecSec, Schweitzer
Electronic Laboratory, Thales e-Security, and Weston Technology.
Peoples Energy (Chicago) and Detroit Edison have also been supportive
and contributed extensively to the working group's understanding of the
needs of SCADA operators.
DESPITE REMAINING WORK, AGA 12 HAS SLOWED SUBSTANTIALLY
Although significant work remains to be done to complete the AGA
standard, progress stopped in May of 2005 when TSWG funding ran out.
TSWG is an organization which only funds prototype developments until
they prove successful, at which time funding is to be provided by other
organizations. DOE has supported Sandia and Pacific Northwest National
Laboratory to evaluate the security level of the standard and the speed
of its encryption, respectively. In October, DOE provided limited
funding for GTI to complete some field testing and write up the
existing version of AGA 12-2 as a document that is in a suitable format
for ballot. This 5 month hiatus significantly reduced the momentum of
the AGA 12 project. Largely as a result of these delays, one of the
three manufacturers that originally committed to produce AGA 12 modules
has stopped work on this project.
Regrettably, AGA 12 became a victim of its own success. Given that
it is well ahead of any other hardware development of cryptographic
protection and manufacturers are developing products, it appears that
market forces have now taken over and there is no further role for
government support.
The apparent success of AGA 12 obscures the additional work that is
required. This includes several topics that--while of great importance
to the success of the AGA 12 effort--are difficult to appreciate. These
include the following:
Conformance testing--While the AGA 12 standard will be
validated by at least two National Laboratories, SCADA system
owners and operators need a ``seal of approval'' to verify that
the particular products they are considering buying actually do
conform to AGA 12 requirements. There is no existing set of
tests that is recognized as providing this assurance.
Next generation design--Because AGA 12, Part 2 is a
retrofit solution for legacy systems, it is the most expensive
and least effective approach to the cryptographic protection to
SCADA systems. Incorporating this protection into products at
the time of manufacture is estimated to be less than half as
costly as adding it after it is in the field. It is critical,
also, that the next generation systems be able to interoperate
with the units that have already had cryptography added.
Large scale pilot test--While the laboratory and
small-scale field tests that have been completed and will be
done in the near future will validate that AGA 12 does work in
the field, this is not a full scale pilot test. Several parts
of AGA 12 that will function well during a small scale test may
prove problematic for larger scale installations. Key
management is a good example. Another is the possibility that
network congestion problems might manifest themselves when many
of the messages are encrypted, but will be invisible in small
scale tests. SCADA operators are more likely to feel confident
in a system that has been tested in a full-scale pilot than in
a system that has only been tested on a small scale.
Key management--Good cryptographic practice requires
that the keys that decrypt the encrypted data and commands be
changed periodically. This ``key management'' must be done
remotely to be cost effective, since the wide geographic extent
of SCADA systems prohibits visiting sites to change keys if a
strike occurs or if an employee leaves.
Forensics and diagnostics--While it is important that
AGA 12 be able to protect SCADA systems from attack, it is also
desirable that these systems detect attacks that are under way,
inform the operator of the attack, and gather possible forensic
information that will facilitate the detection, identification,
arrest, and prosecution of system attackers. Although AGA 12
contains some features that lay foundations for this type of
work, it is far from complete.
Management port--The management port requires some
additional features that are different from those required to
send data and commands.
Coordination of security standards--It is important
that standards groups establish and maintain contact with one
another. There are estimated to be approximately 100 groups
currently developing cyber security related standards. There is
very little contact among these groups, an undesirable
situation likely to lead to duplication of effort and
conflicting standards that no manufacturer will follow.
High speed networks--While AGA 12's early focus on the
protection of low speed legacy SCADA systems is appropriate in
providing protection to the large installed base of these
systems, it is also clear that many of the newer systems will
use higher speed communication links, such as the Internet.
This requires that we be able to maintain as much
interoperability as possible between the low and high speed
networks.
SEVERAL GOVERNMENT STEPS WILL ADVANCE SCADA SECURITY
In summary, we make the following recommendations
Make sure that there is funding for R&D and strong
industry-government partnerships to develop protection of the
Nation's critical infrastructure against cyber attacks.
Progress is being made--the key to moving forward is to
continue R&D efforts and partnerships.
Prevent loss of momentum by avoiding funding
interruptions in on-going programs.
Continue the coordination efforts (such as the DOE
Control Systems Roadmap and the DHS Process Control Systems
Forum) which are key elements of growing coordination between
the government and industry and also vital to coordination
among different infrastructures. These two programs are models
for how to coordinate across a wide area.
Support continued development of AGA 12. In
particular, work should be completed to develop key management,
establish conformance tests, do a large-scale pilot test,
specify a next-generation design, secure high-speed networks in
a manner compatible with the low speed networks, and develop
forensics and diagnostics to detect and foil attacks.
Support selected other standards development efforts.
While our focus here has been on AGA 12, it is important to
recall that this is only a small part of the total SCADA
security requirements. Both the ISA SP99 and the NIST PCSRF
efforts are noteworthy. Many of these other standards groups
labor on an all volunteer basis on other critical requirements
of significance as great as that of AGA 12. This all volunteer
pace will not lead to rapid development of required standards.
Mr. Chairman, we applaud your focus on securing our critical
infrastructure, especially in the area of SCADA protection. This
concludes my prepared statement. I would be pleased to respond to any
questions you or other Members of the Subcommittee may have.
List of Acronyms
AGA--American Gas Association
AGA 12--American Gas Association Report No. 12, ``Cryptographic
Protection of SCADA Communications''
CM--Cryptographic Module
DOE--Department of Energy
EPRI--Electric Power Research Institute
GTI--Gas Technology Institute
ISA--Instrumentation, Systems and Automation Society
ISA SP 99--ISA Special Publication 99, ``Manufacturing and
Control Systems Security
NERC North American Electric Reliability Council
NIST--National Institute of Standards and Technology
PCSRF--Process Control Security Research Forum
RTU--Remote Terminal Unit
SCADA--Supervisory Control And Data Acquisition
SCM--SCADA Cryptographic Module
TNT--Tri-Nitro Toluene (dynamite)
TSWG--Technical Support Working Group, part of the Combating
Terrorism Technology Support Office
[GRAPHIC] [TIFF OMITTED] T2242.001
Mr. Lungren. The Chair would now recognize Mr. Allan
Paller, the Director of Research for the SANS Institute, to
testify.
STATEMENT OF ALAN PALLER, DIRECTOR OF RESEARCH, THE SANS
INSTITUTE
Mr. Paller. Thank you, Mr. Chairman. SANS is different from
the other organizations. We are basically an educational
organization. We--our 45,000 alumni are the front lines, the
people who put the security into the computers that try to
block the attack. So we are constantly looking for methods that
will make that feasible, because right now the bad guys are
winning faster than the good guys are getting better.
So what I am going to do today is not talk about what the
solution is to SCADA security, but how you can take--how we can
prove you can take the solutions that Sam and K.P. and the
others and Bill have found already and get them into operation
rather than studying them to death. So that is what the
testimony will be.
I do want to emphasize that you will sometimes hear these
computers are not connected to the Internet; therefore, they
are safe. The problem with that statement is they are often
connected by packet radio. Think of old-fashioned wireless. So
they might not be on the Internet, but the packet radio is the
method by which the water treatment system in Maroochy Shire
was taken over, and human waste backed up on the streets of the
city, by a man who was angry at the system. It wasn't connected
to the Internet, but it was very vulnerable. So we need to look
at both of those attack methods. And these vulnerabilities
aren't theoretical. You already heard them from Sam.
What I am going try to show you is a method and tell you a
quick story of a method the U.S. Government has used that
radically changed the dynamics of security in the country. And
I think I will tell you that story and then finish this.
Microsoft systems are being put more and more into SCADA
systems. You are buying them. GAO just came without a report
that said that the problem--not just, a few months ago--came
out with a report that says the problems in SCADA security are
getting worse because they are connected to the Internet and
because they are buying off-the-shelf, vulnerable operating
systems.
So how do you make somebody who has a powerful monopoly
over all of the computers that we buy change their way and
deliver safer systems? About 2-1/2 years ago, the CIO at the
Air Force got up at a public meeting and said, we are now
spending more money to fix the problems we have because of
Microsoft bugs than to buy the stuff in the first place. But he
did something that no one else has done. He took Federal
procurement power and said, we are going to fix this. And what
he did is he consolidated all of the contracts that the Air
Force has with Microsoft, all of them, and in doing that he
saved $100 million. It is a half-a-billion-dollar procurement,
but he has got provable savings of $100 million.
But that wasn't the exciting part of it. The exciting part
of it was that he required Microsoft to deliver systems that
were preconfigured according to the standards that DHS helped
create, that the National Security Agency really fronted, and
an organization called the Center for Internet Security brought
together. So there was consensus benchmarks for what safe
means, and that allowed the Air Force to require the vendor to
deliver safer systems. It was a lot of argument, a lot of
negotiation, but in the end Steve Vollmer and Microsoft said
yes.
And what I am trying to show you is you can actually change
the rate at which systems get safer by using combined buying
power, and that is what I believe can be done very quickly in a
SCADA environment, because what Bill is talking about, what Sam
is talking about, what K.P. is talking about are actual
solutions that aren't going to get implemented unless the
buyers can act together, because the vendors--each individual
vendor has an incentive not to get ahead of the others because
it will cost them more, and if they spend more, the other
vendors can sell cheaper. So unless the buyers get together and
agree on standards, it won't happen.
And what is exciting about the SCADA system is the State
and local governments and the Federal Government have a huge
concentration of them, so they can create an enormous buying
power as long as the DHS and Sandia, and Bill and K.P. can
agree on what those standards need to be. And it is a very
quick thing. We are not talking about years and years. We are
talking about weeks and months to agree on what needs to be
done. But then instead of having regulations, instead of having
laws, use procurement power to change things.
I thank you for allowing me to speak, and I look forward to
questions. And I hope you feel better, Ms. Ranking Minority
Member.
[The information follows:]
[GRAPHIC] [TIFF OMITTED] T2242.002
[GRAPHIC] [TIFF OMITTED] T2242.003
[GRAPHIC] [TIFF OMITTED] T2242.004
[GRAPHIC] [TIFF OMITTED] T2242.005
[GRAPHIC] [TIFF OMITTED] T2242.006
[GRAPHIC] [TIFF OMITTED] T2242.007
[GRAPHIC] [TIFF OMITTED] T2242.008
[GRAPHIC] [TIFF OMITTED] T2242.009
Mr. Lungren. I hope she feels better, too. I am not sure I
feel better after hearing your testimony about the
vulnerabilities that we have here.
We have now been informed that I guess we are to go back at
5:30, so we will have time to not only ask questions, but to
hear your comments. And I appreciate your brevity, but I also
appreciate the quality of the testimony.
This is a concern that many of us on this community have.
It is, as someone said, the soft underbelly of our
infrastructure, and it is something that doesn't immediately
come to mind because we take for granted that we have these
systems that work. And our increased interconnectivity is a
blessing, but it is also a curse. It creates the vulnerability
that makes that soft underbelly even greater. And I hope I am
pronouncing your name correctly. Is it Dr. Varnado?
Mr. Varnado. Varnado.
Mr. Lungren. Varnado. I put the wrong emphasis on the
syllable.
Dr. Varnado, of all the things that you suggested are our
vulnerabilities, what would be the chief one; that is, the
greatest--which would require the greater exertion of political
will and governmental attention right now?
Mr. Varnado. There are basically two approaches that we
need to take. We need to continue to work on the inherent
vulnerabilities that are there in every networked computer
system. Industry and universities are doing a pretty good job
in taking, looking at that one very hard.
The second area is that of induced vulnerabilities,
something like what happened in Russia. And the problem with
the COTS products, those are very complex systems. We have no
idea what is deeply buried in those systems. The software that
we purchase may have 20 million lines of code, and for us to
reverse-engineer that is a very difficult task. Same thing with
chips. There can be layers, seven, eight layers, in
microelectronics today. More and more of those systems are
embedded. So finding out how to reverse-engineer some of those
products and to do security checks is a very difficult problem.
Now, the thing that comes to mind for Congress is trying to
improve our collaboration among universities and industry and
national labs and the government. There are things that get in
the way, like classification issues. There are certain things
about the threat that we can't talk about in this room. There
are other issues like trust, antitrust, those kind of things,
that the government could take some action to help give some
relief in those areas so that we could discuss more. If we
could discuss more openly the things that we all know, we would
be in a better position.
So I probably didn't answer your question precisely,
butSec.
Mr. Lungren. Well, let me ask it another way. You said
that--I mean, you almost articulated an insoluble problem which
said we are attempting to build trusted systems with
untrustworthy pieces. Other than us pulling in and saying
everything we are going to do is going to be totally
domestically engineered, produced, testing, et cetera, what do
you suggest?
Mr. Varnado. Doing it all ourselves is not in the cards. We
can't afford it. So what we are doing is we are looking at
different ways to configure systems that put security checks
built into the technology as you assemble the system. So we are
trying to decompose the system a bit and to put in security
features where we think we might find problems and be able to
detect problems quicker.
We do not have intrusion detection systems, for example,
that operate in real time. That is why on the zero day exploits
and the things like the 8 minutes of infecting the DOD system
is so hard. We don't have these real-time intrusion detection
systems yet. So we are working on those kinds of things to try
to solve this problem. We cannot afford to build everything, no
question.
Mr. Lungren. Thank you, Dr. Varnado.
Mr. Purdy, I have had a chance to hear you before, and I am
very impressed with the breadth of your knowledge and the
obligations that you have at your job. Having heard Dr. Varnado
articulate the problem, as well as several other of the members
of the panel, how do you at Homeland Security attempt to try
and deal with this challenge, because in some ways it is a
matter of priorities; and also, how do we--it seems to me that
there is more things you can do immediately by command within
the government than you can do in the private sector. How do
you differentiate between what you can do by command in the
government versus what you can do by whatever means in the
private sector?
Mr. Purdy. Well, that is a difficult question which I know
is one of the reasons that you asked it, and the importance of
trying to get a handle on these issues. But essentially
Secretary Chertoff's approach of risk assessment and risk
mitigation, which underlies our National Infrastructure
Protection Plan, and our work in building the partnership
between government and the private sector for information
technology within the sector and across the sectors, that is a
fundamental piece of our effort. But we have prioritized
several risk mitigation efforts within that context. One is
control systems we have talked about. Another very important
one is software assurance, and a third is Internet disruption,
trying to promote the survivability and resilience of the
Internet.
The software assurance piece that Dr. Varnado talked about
relates to a number of efforts going on that are coordinated.
The Department of Defense has a major effort in the software
assurance area that is closely coordinated with our own
software assurance security program.
The two fundamental things in addition to the purchasing
power issue that Allan Paller talked about which we are working
very hard on is the development of best practices along the
development cycle for software assurance. And it is developing
tools so that we can go back and assess the software after the
fact.
The foreign issue that Dr. Varnado talked about, we are
working in the unclassified and in the classified space. I am
working as the cochairman of the globalization of IT within the
Committee on National Security Systems, where the 24 agencies
are coordinated on the national security systems so we can
address exactly the kind of issue that Dr. Varnado talked
about, the insecurity of what is made overseas, but, in fact,
our inability to be able to tell on what is made domestically
as to whether software not only does what it is supposed to,
but to make sure it doesn't do other things; the coordinated
effort among the partnership, among the national labs; the
funding that DOE, DHS--some direct, as our 15 million that is
up this year for 2006--the money that our Science and
Technology Directorate is funding; the additional funding that
is provided for next year that goes to the I3P program that
Sandia is coordinating; and a number of the specific efforts we
believe are going to in 2006 provide some real deliverables to
help make folks safer. But it is going to continue to require
the partnership among everybody here, the owners and operators
and the security vendors, and it is a difficult and important
challenge.
Mr. Lungren. The Chair recognizes Ms. Sanchez for 5
minutes.
Ms. Sanchez. Thank you, Mr. Chairman.
Mr. Varnado, I wasn't going to ask a question, but you have
me a little curious. When you talked about new systems and then
intercepting them, did you mean like a little systems test as a
piece of that hardware got made, or--I am trying to understand
what you meant as an ability to counteract.
Mr. Varnado. Right. What we are thinking, this is very much
right now, Congresswoman, an R&D project that we are looking
at. If we purchase most of the system and then we put it
together, there are places in the data flow within the computer
system that we may be able to put some small components in that
would detect certain anomalies or violate certain patterns of
use that would alert us more quickly and maybe even be able to
prevent that from happening. So it is very much an R&D project
at this point, and we are just starting to work on it the last
6 months or so. It is brand new. We think it holds some
promise, but it is a huge problem, and we need to put more
effort on it, I guess is the message I want to leave with you.
Ms. Sanchez. Thank you, Doctor.
Mr. Chairman, I am going to give up my time so that--
because you have a lot of Members on my side who showed up to
this, which goes to show just how important most of us think
this is. And I am going to yield back the rest of my time and
move it on.
Mr. Lungren. I thank the gentlewoman for yielding back and
recognize the Chairman of the subcommittee Mr. Reichert for 5
minutes.
Mr. Reichert. Thank you, Mr. Chairman.
Well, I am going to be totally honest. I am coming at this
from a novice's perspective, and so I listened to you. My
background is law enforcement. And so GTI, SCADA, NSTB, SCSC,
NERC, AGA 12, cryptographic module and TSWG, and I had some
more but I will stop. So, book them, Dano, is where I come
from.
I am just really curious, you know, we need to be prepared.
First of all, where are we really today; in your analysis of
where we actually stand today, where are we? Anybody.
Mr. Paller. The demonstrations of vulnerability are active
and scary. So if you want to break into the power systems and
the other systems in the United States, you can hire a bunch of
companies that will demonstrate that it can be done. I just
don't believe that we are at risk of that right away because it
is easier to bring conventional weapons in and blow things up
than to figure out exactly how to use that to blow up a
pipeline. But I don't think we are far away from it, and if we
wait until we see the first strong use of it, there will be no
catching up.
So it is hard to fix a problem when you don't see the
attacks. It is very hard to spend money on that. That is why I
like the Air Force method, because they actually didn't spend
new money. They used old money and the buying power of the old
money to make the change. I don't think there is another way to
do it. There is not on lot of fresh money coming from the
Federal Government.
Mr. Reichert. Thank you. I thought that might be the
answer.
And so when you look at what we need to do to become more
responsive and aware, there is an educational training process
that has to take place, not only some of the things that you
mention in constructing the right system, but people learning
all of the acronyms that I just mentioned, and I am sure there
are a lot more. But how does local--how does the local
government officials, how do they play into that, local law
enforcement and also the local businesses?
You touched a little bit upon the industry and how they
play a partnership, but when it comes to training, I think, Mr.
Purdy, you mentioned training, and, Dr. Ananth, you said
something about training 350 people. What kind of training, and
who do you train?
Mr. Ananth. Well, if I might say, the training that we talk
about is for the people who install those control systems and
for the end user. We are not talking about training the State
and local people, because, as you know, sir, there is a lot of
problem in interoperability devices with the response workers
and the emergency response workers. But what we are talking
about is the people who actually own the critical
infrastructure assets, which is a lot of the private sector. So
we are talking about where the control systems are located, so
they need to know where the vulnerabilities are, they need to
do a fix. So when we talk about the training, and when I talk
about the training, that is the audience, the target audience,
I was talking about, the owners of these infrastructure assets.
Mr. Reichert. Mr. Purdy, did you have any comment on the
training?
Mr. Purdy. Well, we have a number of different levels of
the awareness piece that was touched on. I believe the House
passed today a resolution to support National Cybersecurity
Awareness Month, which is October, which helps emphasize the
importance of getting the cybersecurity important message out
to consumers and small business and what folks need do about
it.
In addition the training program, we work with the National
Science Foundation on the Cyber Corps Program, because we want
to encourage the number of well-trained cybersecurity
professionals in the Federal workforce, and as part of training
we have been partnering with the Department of Defense, because
one of the big issues about whether the Federal Government has
enough well-qualified people is, if you define all the jobs
differently, it is impossible to do the gap analysis. So they
have done the job task analysis of DOD, and we are going to try
to leverage that across the Federal agencies.
Also we are partnering with the National Security Agency.
In fact, we have a major conference tomorrow up in Baltimore
with the Centers of Academic Excellence, as we have been
creating a common body of knowledge for those university
centers of excellence to train the next generation of
cybersecurity professionals and software developers to do a
better job of what it is that they do.
Mr. Reichert. Thank you, Mr. Chairman.
Mr. Lungren. Mr. Purdy, I might just mention your reference
to the bill that we passed yesterday. It is a great analogy for
where we are. We passed appreciation for this month in the
middle of the month. Maybe it shows you how we have to catch up
in this whole arena.
Mr. Pascrell is recognized for 5 minutes.
Mr. Pascrell. Mr. Purdy, I want to start off with this
question, and I would ask you to be very direct and specific.
How many Department of Homeland Security employees are
currently working on the SCADA control systems issues? How many
people?
Mr. Purdy. We have two government employees and 35 full-
time contractors.
Mr. Pascrell. So there are only two people in the
Department of Homeland Security, and listening to the
vulnerabilities from you six gentleman, we have two employees,
Federal employees, and we are contracting out most of this
work, correct? Correct me if I am wrong so far.
Mr. Purdy. On the control systems piece. The other efforts
we are doing will help protect the control systems owners and
operators as well, and that is integral to it.
Mr. Pascrell. Well, then, let me ask you this question. We
saw in the recent hurricane, Hurricane Katrina, that the
Federal Government was unprepared to respond to a large natural
disaster. Today we have heard about the devastation that may be
caused if a terrorist or a--or there is a natural disaster hits
our control systems. Mr. Varnado, you made four very specific
recommendations. Just last week there was a headline in the New
York Times that said, U.S. Cybersecurity Due for FEMA-Like
Calamity. Are we prepared for a cyberattack on our control
systems, Mr. Purdy? And if a natural disaster hits our control
systems, are we prepared to respond to it, in your estimation?
Mr. Purdy. Well, we believe we are prepared for a
cyberattack, to respond to a cyberattack against the control
systems. Our partner division within the Infrastructure
Protection Office, Protective Security Division, is the best
division to talk about the actual direct physical consequences
of your question.
Mr. Pascrell. So from your standpoint we are prepared.
Mr. Purdy. We have a high cyber risk in this area, but we
are prepared to respond and mitigate an attack that might
occur, yes, sir.
Mr. Pascrell. Well, there is no need to get on the
defensive. I have a right to ask the questions, and you have a
right to deliberate before you answer me.
I am getting particularly annoyed--for the Chair, I am
getting particularly annoyed with employees that come here from
the Department of Homeland Security, the responses to this
committee or any committee dealing with homeland security, and
frankly, I am tired of it because we are not prepared. You know
it, and I know it.
And let me make some suggestions before I leave it for now.
We know that there are vulnerabilities within these systems,
and we know that these vulnerabilities are abundant, and we
know that the threat of the terrorist attack against these
systems is real. Those things we know, we agree on. So the
Congress, it would seem to me, needs to engage in a robust
analysis and oversight in this realm, Mr. Chairman. We need to
help ensure the security of the various control systems that
are used in critical infrastructure. And I am heartened that
today two Homeland Security subcommittees are leading the
charge.
A cyberattack on one of New Jersey's four nuclear power
plants or 100 chemical sites, for example, has the potential to
be absolutely devastating not only in terms of lives lost, but
also in the regional and national economic structure it could
bring forth. That is very serious, very serious business.
Back in 2002, the National Infrastructure Protection Center
reported that a computer belonging to an individual who had
links to Osama bin Laden contained programs that clearly showed
the individual's interest in the structural engineering of
various critical infrastructures. It indicated that al-Qa'ida
members had sought information about the control systems which
we are talking about here today, from the verySec. from the
many multiple Websites.
The NIPP, the National Infrastructure Protection Plan, was
due in December of 2004. Mr. Chairman, please hear me on this.
This is important. The American people, American public is
being duped. That was supposed to be completed in December of
04. In February of 2005, we had an interim plan. It was issued,
setting a deadline of November 05 for the final plan. Now,
according to the GAO, the interim plan was incomplete in the
first place. It lacked both national-level milestones and
sector-specific security plans. The plan remains uncomplete to
this day. We can't even get proposals ready in a timely manner.
This is unconscionable. There really is only one full-time
employee staffed in the DHS that deals with national
cybersecurity, and I am not going to accept as a Member,
Ranking Member, Ranking Member, it doesn't matter, I am not
going to accept folks coming before us and thinking that we
don't do our homework. And we are saying--we are talking here
about on a nonpartisan basis.
This is critical stuff. You have never met deadlines. You
don't care about those deadlines, and I don't think you have
the expertise to meet the deadlines. What do you know about
that? And I have not heard anything to the contradiction to
that statement either. And I am tired of it, and the American
people are tired of it.
Natural disasters. We are not going to have 7 days to
prepare for a terrorist. We are not going to have 7 days. I
suggest that you look at, if you haven't already, Mr. Varnado's
four recommendations. It is a start. It is not the total
solution. There is no seamlessness in this battle, no perfect
systems, but it is 4 years later, and we are no further down
the line, Mr. Chairman.
Thank you for your tolerance
Mr. Lungren. The gentleman's time has expired.
Mr. Purdy, if you wish to respond, you may.
Mr. Purdy. I expect that when the National Infrastructure
Protection Plan goes out early in the year, that the concerns
raised in the GAO report will be well addressed. The work we
have done in the National Cyber Security Division to implement
our strategic plan in furtherance of the national strategy to
secure cyberspace, we believe, has made concerted progress. It
has been reflected in the additional funding we have been
given.
We believe Secretary Chertoff believes in the importance of
the cyber issue as part of the overall risk management
framework that he has. We are proud of the progress we have
made. We would be happy to brief the Congressman and his staff
and other members of the committee on that substantial
progress. I recognize that the cyber risk is substantial. We
recognize it is substantial. We agree with the committee. We
agree with the members of the panel on that issue. To the
extent the forcefulness of my answers came across as being
defensive, I apologize, but that is how forceful I am. Thank
you, sir.
Mr. Lungren. Thank you.
Before I recognize Mr. Pearce, I might just say there has
been some frustration exhibited by this panel for the failure
of reports to be done in a timely fashion, and I think that has
been on a bipartisan basis. There is no suggestion on my part
that you are not trying to do your job, but I will just tell
you that is a real frustration on this committee.
Mr. Pearce is recognized.
Mr. Pearce. Thank you, Mr. Chairman. I have got several
questions, so I am requesting briefer answers if you could.
Mr. Purdy, can you outline the process by which the four
components, divisions of the Office of Infrastructure
Protection coordinate and share information in the progress or
implementation of your mission? You have got four divisions.
How do you all coordinate and share information?
Mr. Purdy. You are talking about generally?
Mr. Pearce. Generally, yes.
Mr. Purdy. Across the board, well, we have two meetings a
week with the Assistant Secretary for Infrastructure
Protection, each of the division directors. We have an
additional meeting without the Assistant Secretary where the
division directors themselves come together. We have milestones
that come down from the Infrastructure Protection Office
weekly. We have weekly reports that the Infrastructure
Protection Office gives to each of the divisions so that people
know what the other groups are doing. And we have a number of
specific areas that we are partnered with; for example, the
Protective Security Division, they do the site-assist visits of
the localities, and we provide the cyber guidance for those
assessments that are due in the local locations. In addition,
we have periodic briefings, where each division briefs the
entire group, all the division heads from Infrastructure
Protection, as to what the goals, objectives, accomplishments,
budgetary situations are, progress and challenges ahead
Mr. Pearce. Mr. Todd, do you have--you have heard Mr.
Purdy's discussion. In your report you talk about the need in
the future for maybe coordinated contact with other agencies.
In the past year what contact have you had with Mr. Purdy's
National Cyber Security Division?
Mr. Todd. Well, let me handle them in two different ways.
One is the--
Mr. Pearce. If you could just give me the brief answer.
What contact have you had with them?
Mr. Todd. I have not had any with them.
Mr. Pearce. Thank you.
Dr. Varnado, what contact has your group had?
Mr. Varnado. We are currently working with him on the
National SCADA Test Bed as well as a program at Dartmouth that
we are interacting on.
Mr. Pearce. Okay. Thank you.
Mr. Todd, as I read your report, I just find the language
to be very reassuring, very reassuring, and I find the language
of the other reports to be not so reassuring; that is, I hear
pointed comments. In other words, you say that you all have
made the appropriate improvement measures, engineering, that
you have done what you can to protect the equipment and ensure
the safety of public health, that you have maintained a policy
of not connecting your SCADA systems. You have evaluated and
improved, you have identified the cyber vulnerabilities. You
are continuously evaluating. Now you list a couple of sections,
but then your closing statement says that we believe our
security program meets the challenges of these requirements,
and then kind of a throwaway comment that we will look forward
to contributing and just staying on top of the situation.
Do you find the reports of the other agencies, the other
people testifying here today, to be that much different from
your findings? In other words, I find some element of alarm in
everyone else's, but yours declares that we are on top of it,
and we have been on top of it, and we are going to stay on top
of it.
Mr. Todd. Well, let me say it this way; the differences, I
believe, are this--we are an agency that puts things out on the
ground. So we are certainly vulnerable to the kinds of
contractors and chips and so forth that we might contract for.
That is true. However, in our implementing these kinds of SCADA
systems, we have had, over the last 20 years, a basic distrust
of the system itself. We want it to be foolproof. And so we
have put in other kinds of guarding devices. For instance, we
have operators on 24 hours a day. We check with transmission
agencies continually about what is being provided and what
isn't. And if those things are not right within our
parameters--
Mr. Pearce. You feel like you could fight off any attempts,
like the Australian attempt that is reported by one of the
other presenters, that that really would not happen in your
agency, that there is not much attempt or much capability for
an outside group to come in and affect the flow of waters
through the BOR or through the dam system or--you know, you
think that you really are that secure.
Mr. Todd. We believe the risk is low.
Mr. Pearce. Okay. Thank you. Appreciate it, Mr. Chairman.
Mr. Lungren. I thank the gentleman.
As I understand it, you do not have the SCADA systems
running the gates; is that correct?
Mr. Todd. We do not have SCADA systems running spillway
gates. We certainly have them running the smaller power gates
for power generation, that is true.
Mr. Lungren. But the greater danger is with the spillway
gates.
Mr. Todd. Yes, it is. Our SCADA systems are set to operate
within the safe channel capacity, and so, therefore, we do not
have them hooked up to the spillway gates, which are set to
operate sometimes out of the channel capacity.
Mr. Lungren. I thank the gentleman.
Mr. Lungren. It is my pleasure to recognize the chairman of
the full committee, Mr. Thompson.
Mr. Thompson. Thank you. I am interested to know from Mr.
Pearce's answer that it was low risk. And the chairman just
asked the question--you said, it was high risk; if I could get
clarification and communication from one to the other, with the
dams.
Mr. Todd. Excuse me, I am not quite understanding the
difference of what you are asking.
Mr. Lungren. Mr. Chairman, I was asking about--the highest
risk, as I understand it, comes from the control of the gates
from the spillways and they are not on a SCADA system. Even
though they have a SCADA system that does deal with the gates
that go to the power plants, it deals with the volume, so the
highest risk.
Mr. Todd. Okay, I think I understand what you are asking.
Our SCADA systems operate power plants, and in those power
generation plants they have turbines which--we have special
inlets which have some gates to those turbines. Those are much
smaller systems that, if all were turned on, for instance, full
speed, they would still operate within the channel capacity
downstream, so it wouldn't cause a catastrophe or consequences
of damage and that sort of thing.
However, we also, in operating the dam, have much larger
gates because of high flooding and other kinds of events that
we have to safeguard the dam itself. Those gates, which if
operated at full capacity, might go out of the channel
capacity; those gates are not hooked up to the SCADA systems.
So our SCADA systems would only operate within the safe channel
capacity, itself, of the river.
Mr. Thompson. Is there a plan to put them on the system?
Mr. Todd. Not that I am aware of.
Mr. Thompson. Mr. Purdy, the President asked in 2003 that
we put together this National Infrastructure Protection Plan.
As you know, we have more or less missed deadlines, and when we
finally got it, GAO was very critical of the product. It was
pulled back, and I would assume that at some point we will have
another response or report put together.
Do you have any idea when we will have that?
Mr. Purdy. Well, I will expect the report to come out
shortly after the first of the year. Once that report comes
out, then the sector-specific plan--such as, our sector is
information technology--there will be a 6-month period in which
we work with the private sector to create those plans.
So the specific implementation plans in each sector will be
ready 6 months after that.
Mr. Thompson. So we will miss the November deadline?
Mr. Purdy. Well, I will leave that up to my boss, the
Assistant Secretary, to--I believe he is coming to the Hill on
Thursday. So I probably shouldn't officially comment on meeting
that deadline, but I am confident it will be there shortly
after the first of the year.
Mr. Thompson. Okay. All right.
Well, Mr. Chairman, I hope you noticed that we are still a
little tardy with our deadline.
Mr. Lungren. I understand that. I also apologize for
calling you chairman. Either--
Mr. Thompson. No, I accept.
Mr. Lungren. Either I have granted Ms. Pelosi's fondest
wish or I have inducted you into the Republican Hall of Fame,
so whichever one you would like.
Mr. Thompson. Well, okay, either way, I accept.
The other thing, Mr. Purdy, I am a little concerned about
is the fact that we don't have but two full-time employees in
your Department; is that correct?
Mr. Purdy. We have two Federal employees working on the
control systems area and 35 contractors. We have an allocation
of 40.
Mr. Thompson. Explain the contractors to me.
Mr. Purdy. They are people paid--many of them are through
the national labs, for example, people that are not official
government employees that are paid on a contract basis through
a contractor. That is supporting our efforts in the control
systems area. My division is the National Cyber Security
Division which--control systems is one part of a broader
effort.
So we have an allocation of 40 Federal employees. Of those,
we have two and one to be hired for the control systems area
that are official government employees.
Mr. Thompson. Now, the contractors, are those individuals
that are contracted?
Mr. Purdy. No, they are through companies or through the
national labs.
Mr. Thompson. All right.
Can you provide this committee with how much that is
costing taxpayers, rather than having full-time employees, how
much we are paying those contract employees?
Mr. Purdy. Yes. We can get you how the funding is broken
down by contractors, yes, sir. We can get you that.
Mr. Thompson. For the record, can you tell me whether or
not we are paying more for those people based on contracts than
if they were full-time employees?
Mr. Purdy. I can't. I haven't seen the per-person breakdown
of it. So I can't answer that question, sir, but we will be
able to give you information from which that will be obvious.
Mr. Thompson. Well, just tell me what your best guess is.
You are over it, right?
Mr. Purdy. I couldn't hear you.
Mr. Thompson. You are over it, right, you are over the
division?
Mr. Purdy. Yes.
Mr. Thompson. You approve the contracts?
Mr. Purdy. Yes.
Mr. Thompson. Well, just give me a best guess whether we
pay more for the contract employees rather than if they were on
a Federal payroll.
Mr. Purdy. My best guess is, we are paying more for
contract employees, yes, sir.
Mr. Thompson. How much more?
Mr. Purdy. Sir, that really would be a guess. I really
shouldn't venture there.
Mr. Thompson. Well, so is it your opinion that we get a
better product with contract employees than full-time
employees?
Mr. Purdy. I am given a certain allocation of Federal
employees to achieve our mission and implement our objectives
and goals. So to do that, we need to hire contractors to help
us fulfill our mission.
Mr. Thompson. So, in other words, you can't hire but three
people?
Mr. Purdy. We can't hire but 40 people, right.
Mr. Thompson. Out of that 40, you chose to hire three in
the Federal system and then contract everyone else?
Mr. Purdy. That is correct.
Mr. Thompson. Even though it costs us more to contract,
there are 37 others?
Mr. Purdy. Yes, sir.
Mr. Thompson. Well, I guess when you have got a lot of
money, you can do that.
Thank you, Mr. Chairman.
Mr. Lungren. Thank you, Mr. Thompson, Ranking Member
Thompson.
Ms. Brown-Waite.
Ms. Brown-Waite. Thank you very much, Mr. Chairman.
Perhaps as a follow-up to Mr. Thompson's question, if these
are individuals and you have a contract with them, you
obviously have a deliverable. What are they supposed to be
delivering?
Mr. Purdy. I am sorry.
What are the deliverables? Maybe that would help us to
understand, when you do respond to the question, if you would
also put what the deliverables are, because it could very well
be that there isn't any qualified employee.
I think, in addition to the deliverables, a natural follow-
up question is, what is the length of their contract and when
are they supposed to produce and what are they supposed to
produce? I think that would be very appropriate.
I know you probably don't have that with you now. But in
addition to how much are we spending, I think that that is an
important follow-up component.
The SCADA system is something that I was familiar with. I
used to be a contracts manager at a water management district,
which meant I got to okay the payments, the monthly and
quarterly payments for the SCADA systems, for their structures,
their control structures. So, naturally there is a concern, you
want to make sure that they work. But that was long before 9/
11, so when you look at all the other systems, obviously the
whole SCADA system of controls is just very, very important.
While we have concentrated on how many employees work for
you on SCADA, maybe we also need to ask, do you know how many
are at NCSD?
Mr. Purdy. Well, as I said, we have an allocation of 40. We
have 25 or 26 in place. I believe we have six or seven in the
hiring pipeline; we are pursuing hiring an additional balance
of the 40.
Ms. Brown-Waite. Okay, and in a follow-up question, what is
the plan for the NCSD in the reorganization?
Mr. Purdy. Our division will move, of course, into the
larger preparedness directorate, the information analysis,
infrastructure protection directorate; that is, the Under
Secretary level has become a preparedness directorate.
Within that, we will move along with the telecommunications
folks, called NCS, National Communications System. So cyber and
telecommunications will be under a new position that is being
created for an assistant secretary for cyber and security
telecommunications. So we will be under a new assistant
secretary who will, in turn, be under the under secretary for
preparedness.
Ms. Brown-Waite. I can tell you that so many constituents
just feel that the Department of homeland Security is nothing
other than bureaucracy, layer upon layer, and that there is
just a lot of concern out there that the major question is, are
we safer for it today.
Can you also tell me, Mr. Purdy, what progress is actually
being made in developing standards for SCADA systems?
Mr. Purdy. Well, some of the members of the national labs
here can probably go into more detail than I can. But within
the framework of our plan for 2006, there was some discussion
about the cyber security protection framework to develop and
disseminate tools to assist the users in assessing their cyber
security practices against industry best practices and
standards. We are trying to work to perform those vulnerability
assessments to identify the weaknesses in the systems against
those standards and recommend mitigative strategies for them.
The Process Control Systems Forum, which we cosponsor with
the Science and Technology Directorate, with the users--again,
we are working with the owners and operators, the vendors and
the national labs to help identify the specific standards for
the control systems against which we can judge how the actual
owners and operators are doing.
Ms. Brown-Waite. So I think what you said is, there is no
standard yet, but you are working on it. Is that--
Mr. Purdy. We have a draft cyber security framework, as I
said in my testimony, that we are going to be piloting this
year, that we will then be able to roll out this year--``this
year'' being 2006--so that the individual companies can do
their assessments. That is going to be part of the effort as
discussed by others to build the business case to convince the
owners and operators to spend the money to meet the standards.
Ms. Brown-Waite. Do you believe that there is a way that
government can incentivize the private sector to actually
develop smarter SCADA security?
Mr. Purdy. Well, within the context of software and in
controlled systems, we want to do--and we have begun to do what
Alan Paller was talking about, which is put in incentivizing
programs for those contracts which the Federal Government is
buying so that we can raise the bar in a nonmandatory way--not
like in a regulatory way, but if you want to get the contract,
you have to have the security built into the system you are
selling. We believe that is an important basis.
In addition, having the assessments and the framework and
the tools for self-assessment, that is going to help encourage
and make the business case for the private sector to spend the
money.
Ms. Brown-Waite. Thank you, Mr. Purdy.
I yield back my time. I thank the chairman.
Mr. Lungren. I thank the gentlelady.
The gentlelady, Ms. Norton, is recognized for 5 minutes.
Ms. Norton. Thank you, Mr. Chairman. Actually, it is this
latter point, and I was going to direct the question to Mr.
Paller, because I was intrigued with his notion of requirements
of the contractor essentially to deliver security-ready
systems.
It seems so obvious that I have to ask you--it seems
obvious because, obviously, if you are delivering to the big
granddaddy of them all, the Federal Government, you really do
call the shots. You know, it is like Texas calling the shots on
textbooks that everybody else has got to use, because they have
more kids. Or it is like Medicaid prescription drugs, where we
ought to be taking advantage of our market advantage. This goes
to the underlying substance: Who in the hell needs this more
than the Federal Government?
What are the--I mean, what do we--what are the barriers? I
mean, for example, is this very costly to do? If so, you know,
I can't imagine that it would cost us even more to do it after
we got it. So that is one question.
Are there security reasons? Is there some discussion of
contractors and whether or not you want them that much, excuse
me, in on your business, but they, of course, I presume, know
all this in the first place.
I would like to know what are the real barriers to this and
whether it can be done, because you indicated it can be done
pretty quickly.
Mr. Paller. There are two barriers that we have seen, one--
and they are both real, so that when people fight against it,
they are fighting not irrationally.
One is, if you take responsibility for securing systems and
you deliver a more secure system, when the user wants to do
something that is not turned on by default, he may call up for
support. So there is a support issue that comes in.
But the much larger one, that the lawyers get involved in,
is that they are worried about taking liability. They are
concerned that if they say, now we are going to give you a more
secure system, that somehow the trial lawyers will be all
around them. At least that is what they say.
But could I just take one second and answer another
question that I wasn't asked?
Ms. Norton. On my time?
Mr. Paller. Yes.
Ms. Norton. No. Because I have another question.
Mr. Paller. All right. You.
Mr. Lungren. Normally, we would allow you, but we have a
short time frame here.
Ms. Norton. If he would have extended my time--see, he is
not going to do that.
I have got to go to Mr. Purdy and ask him about the four
cyber security managers in so short a period of time, high
turnover, and of all positions, the security managers at DHS.
As I understand it, the last turnover was in January. This
doesn't make me feel very secure.
Mr. Purdy, I would like to know why there is such turnover
in the cyber security managers, what you can do to correct it.
I can't believe it is good for the system.
I want to know what the effect is on cyber security, and I
want to know why the Secretary hasn't appointed a new cyber
security manager here in the month of October?
Mr. Purdy. Well, let me address the last question first. It
is certainly my expectation and hope that now that the new
directorate is stood up by the President signing the Department
of Homeland Security budget, that Secretary Chertoff will
announce the appointment of an assistant secretary for cyber
security and intelligence.
Ms. Norton. Excuse me, so you are saying it was a budget
question?
Mr. Purdy. The position did not exist before the President
signed the budget. All I am saying is, it is my expectation.
Ms. Norton. I thought there were four cyber security
managers. So you are saying the position of cyber security
managers did not exist?
Mr. Purdy. I am sorry. I am trying to answer your last
question first, the question on Secretary Chertoff appointing a
new assistant secretary for cyber security and
telecommunication. And I was saying, it is my hope and
expectation that he will make that appointment very soon now
that the new directorate has been stood up.
I think the publicity about high-level departures from my
division is really overblown. To me, the progress that we made
from the time I came over from the White House, having worked
on the national strategy, in April of 2003, through the time
when Amit Yoran, my predecessor, was in office and some of the
other folks have left and are gone, we have built and have
implemented a very important complex plan to reduce our cyber
risk. We do not believe that has been impacted by individuals'
departing.
Ms. Norton. Why are they departing? Please answer my
question; I have limited time.
If there have been these rapid departures, one, why have
they departed; and two, what can we do to keep turnover in all
divisions of cyber security managers? I would like to ask my
question because, you know, everybody is going to leave here in
a minute.
Mr. Purdy. Some of the positions were departures based on
personal reasons that were not related to mission. I think that
is primarily what we are talking about, not related to mission.
We believe we have the positions in place. We have the plan
in place. We have the funding, particularly with the additional
2006 money, that we are going to be able to keep strong people,
and we are going to be able to implement our strategic plan.
Ms. Norton. I will accept that as a promise.
Thank you, Mr. Chairman.
Mr. Lungren. I thank the gentlelady.
Ms. Jackson-Lee is recognized for 5 minutes.
Ms. Jackson-Lee. Thank you very much, Mr. Chairman, and to
the ranking member. We don't have a lot of time for what I
think is a very important hearing.
I guess I remain troubled by, one--Mr. Purdy, maybe you can
tell me, you might be under review or under the consent process
of the Senate. You might advise me of that. But I continue to
be troubled by the acting director scenario, because I think in
the Department of Homeland Security we are rattled, if you
will, with interim and acting personnel when we have a very
serious challenge. So I know in the course of your response,
you will provide me with that.
I would like, first of all, to ask unanimous consent to
have my statement submitted into the record, Mr. Chairman.
Mr. Lungren. It is so submitted.
Ms. Jackson-Lee. But what I would like you to walk me
through again, and if you have said this previously, thank you
for repeating it in a more detailed manner, and that is the
absence of a National Infrastructure Protection Plan. Why don't
you tell me why no such plan exists?
I am sure you are going to tell me that it is either being
worked on or it has been submitted, and I missed it. But then
also tell me what you would expect to see in such a plan?
Let me just highlight for you that in the course of at
least 6 months, we have had a number of incidents at our
chemical plants and refineries in the gulf coast region. Adding
to the misery, of course, were Hurricane Katrina and Rita in
terms of control data systems determining the status of those
particular entities, one, the vulnerability to terrorism and
other catastrophes that might make the situation worse.
So obviously this hearing is extremely important, because
we are talking about control systems and SCADA systems which
are sometimes confused and intermingled.
I think it is obviously a failure that we have never
finished our national vulnerability assessment or national
threat assessment that I think many of us have been asking for
for a number of years now, since 9/11.
Now I understand that we don't have the particular National
Infrastructure Protection Plan relevant to the issues at hand.
Would you, first of all, respond to--you could give me your
status, but would you both tell me whether there is an existing
plan, but then what you would expect or would see, expect to
see, in such a plan to be presented and to be in place?
Mr. Purdy.
Mr. Purdy. So the existing status, you are talking about my
acting director position?
Ms. Jackson-Lee. I am. Are you acting or are you in the
middle of being confirmed?
Mr. Purdy. No.
Ms. Jackson-Lee. Or what is your stance?
Mr. Purdy. No. I am the Acting Director of the National
Cyber Security Division, and we are waiting for the appointment
of an assistant secretary for cyber security and
telecommunications, who will be my boss; and he or she will
make the decision of whether I will be director or in some
other position.
Ms. Jackson-Lee. So we are in complete disarray?
Mr. Purdy. No, I think we are implementing our strategic
plan in furtherance of the National Strategy to Secure
Cyberspace. I think we are making demonstrable progress, and we
are happy to brief you in more detail on it.
Ms. Jackson-Lee. Can you help me then with the question
that I asked, why do we not have such a plan right now?
Mr. Purdy. The responsibility for the plan is the
responsibility of my boss, the Assistant Secretary.
Ms. Jackson-Lee. Who doesn't exist at this time?
Mr. Purdy. The Assistant Secretary for Infrastructure
Protection, until the time that President Bush signed the
budget, was my boss. When the budget is signed, as soon as my
bosses tell me that there is a change, then there is a vacancy
creating an assistant secretary for cyber security and
telecommunications who will be my boss. So we are in a little
bit of a transition period.
But in response to your question, they didn't want to make
a decision to drop the ``Acting'' from my title, giving the
opportunity to the person who will be my boss, so that he or
she can decide who they want in that position and how they want
to organize cyber security and telecommunications in a cohesive
and integrated way.
Ms. Jackson-Lee. Let me acknowledge that I am putting you
in probably an untenable and embarrassing and compromising
position in terms of trying to answer the question. Let me
thank you, first of all, for your service, but let me admit
that what you have just said sounded as convoluted as one might
imagine.
It is almost incomprehensible what you just said. I think I
gleaned from it that someone that was in the position went on
to something else, and they are dealing with the budget, and
therefore, we are not in order.
I would only say to you this: The acts of terror really
don't make appointments, and they don't respond to our lack of
personnel in place. So your response certainly is not your
responsibility and fault. But let me go on record and say that
we are in disarray, and we are dangerously in disarray in a
very important area.
I do acknowledge that recent legislation had funding in the
cyber security area, and I am very glad of that, and amendments
that we have put forward have been accepted, but still--would
you please answer the question again?
I don't think we will agree on whether or not the area
where you are in is in order, but can we at least agree, is
there or is there not a National Infrastructure Protection
Plan, yes or no; and if there is not, prospectively what would
you expect to be included in that plan?
Mr. Purdy. The draft of the National Infrastructure
Protection Plan is on Secretary Chertoff's desk, and we expect
it to be circulated for additional comment in the next few
weeks.
Mr. Lungren. The gentlelady's time has expired.
Ms. Jackson-Lee. I thank the chairman.
Mr. Lungren. The gentleman from North Carolina, Mr.
Etheridge, is recognized for 5 minutes.
Mr. Etheridge.7 Thank you, Mr. Chairman.
Mr. Purdy, at the risk of embarrassment, I am going to go
back to the issue that we are still on, and then I have--I am
going to move on and try to get to another question.
As you draft the response to this question on the budget
that you had indicated you will share with us relative to the
40 slots that are available in your area, I recognize that you
are only the Acting Director. But that doesn't matter; this
committee deserves the information.
I would like to know, and I think the other members of the
committee would like to know, as you look at that, since we
only have three permanent positions, what--as you draft the
numbers for the cost of the contractors, how much the taxpayers
of this country would be saving if we had full-time positions
and what the turnover would be if they were not contractors
that moved back and forth.
I think it is critical--and I am going not going to ask you
to answer that today, but I think it is a critical issue to
have permanent people you can have access to, that can be
trained, who aren't likely to have the information and you have
to move on and you have to have different people in place. I
think that has a real impact on continuity.
Because you said early on that cyber security is important.
I happen to believe it is, and if it is important, it ought to
be important enough to have permanent, full-time people to be
there in place on a daily basis to deal with these issues that
are important to the taxpayers of this country and to the
people of America.
I hope you agree with that.
Mr. Purdy. Yes, sir.
Mr. Etheridge. I hope you will add that to the material you
are going to send us.
Now, my question is this: I wanted to follow up, and you
probably can't answer it, because you have tried to get to it
and haven't really answered it thus far, simply because I think
it is above your pay grade, and that is inappropriate, because
having as many people in this position since the Department has
been funded creates a real problem of continuity for people
now, in this period of time, without having someone permanent.
I am going to leave that out there and not ask you to
respond to it, because I think it is inappropriate to ask you
to respond to it. But I trust this information will get back to
the Department. Hopefully, the Secretary will be here at some
point, and we will have an opportunity to ask that question.
My question to you and to Dr. Rush and Mr. Paller--I will
say this: The Department of Homeland Security established the
Process Control Systems Forum to facilitate communication
between government, industry, vendors and academia. Are you
familiar with that?
Okay.
How effective has this endeavor been, and do you know of
any meetings between these groups? If you do, what was the
outcome?
Mr. Rush. Yes. I would say those are some of the most
effective activities I have seen.
We are developing standards; we are feeding them in. There
are two activities--well, really three, but the two that you
mentioned, the PCSF, the Process Control Systems Forum has
brought together the vendors, the manufacturers, the users,
cryptographic experts, the whole field. That has been very
effective.
There was a question about coordinating Chairs. We had a
meeting just a couple of weeks ago where there were literally
dozens of organizations getting together and swapping
glossaries and making substantial progress.
Mr. Etheridge. Beyond philosophies, though, did we get any
results?
Mr. Rush. Absolutely.
Mr. Etheridge. Can you name, share with us some of the
results?
Mr. Rush. In terms of things that are out there?
Mr. Etheridge. Yes, please.
Mr. Rush. Here is a product that conforms to one of the
standards. What you need to understand is the standards groups
are volunteer organizations, and they don't have the resources
to coordinate. This provides them with exactly the forum that
they need to exchange. We have got 100 groups working
independently. Imagine 100 congressional committees not talking
to each other.
Mr. Etheridge. Good. Thank you.
Mr. Paller. Yes. It is a wonderful talking group. Bill's
outcome is very real. There is a problem with groups like that.
It was seen in the health--the security of the health devices,
CAT scanners and things like that.
When the vendors have too big a role, implementation of
security is delayed almost endlessly. So at some point, the
users have to say, this is our need, our things are at risk.
Vendors are going to have to deliver what we say rather than
letting the vendors hold it up.
So PCSF is the best thing out there, but at some point the
vendors will have to be asked to wait outside while they vote.
Mr. Etheridge. Mr. Purdy.
Mr. Purdy. In addition, the PCSF has provided the input
that has led to the development of the security framework,
which helped set the best practices and also provided the input
for the development of the assessment tool. The assessment
tool, which is now being used to test, is used to assess the
cyber components of the control systems and then provide the
checklist and the questionnaire to determine the particular
vulnerabilities and whether the mitigated steps have been put
in place. That collaborative effort is what is helping to drive
solutions to a very complex problem.
One of the reasons for the complexity is that so many
different owners and operators have so many different systems
with different levels of maturity. So it is hard to have one
set fix across the board to make it better. So that is why the
collaboration in developing these tools in the framework has
been so important.
Mr. Etheridge. Thank you. I yield back.
Mr. Lungren. The gentlelady from the Virgin Islands, Mrs.
Christensen, is recognized for 5 minutes.
Mrs. Christensen. Thanks, Mr. Chairman. Let me ask a little
bit different question.
I want to ask Mr. Paller about the training, because that
is your responsibility also, the training of the technical
security professionals. Where are we, how many have you
trained? What is our need? How are we meeting that need?
Also, where did the students come from? And do you work
with universities, and if you work with universities, to what
extent are minority-serving institutions involved?
Mr. Paller. When we get all done training everybody we can
train, we won't have touched 1 percent of the people who have
control of these systems. So the solution is not to train more
people. We have got to build safer systems; then the training
will have an effect. So as hard as we work, we will never get
there.
I do want to go back to Mr. Reichert's question. We
actually work with universities and local law enforcement. They
don't have the funds that large companies do, so we have major
programs where we cut the costs of education by about 85
percent, so they get a much lower cost. So locally we work with
the FBI to set up these programs for local law enforcement. It
actually is wonderful, because they give more feedback, and
they are the best students we get.
But the training of the SCADA people, we have just begun
with courses on how you measure SCADA security, and they are
just starting. I think the jury is still out. You have got two
groups. You have SCADA engineers on the one side and security
people on the other side, and getting the course right for
those two interest groups is challenging. So we will know in
the spring how that works. .
Mrs. Christensen. Okay, just one other question for
Mr. Todd. Since I sit on the Resources Committee, I am glad
to know that your SCADA system is not connected to the
administrative systems because that is one of the problems we
are reading about.
Do you monitor only the 17 dams that the Bureau has created
or are you monitoring the private dams? Have you used the RAM-D
to assess the threats, vulnerabilities and consequences; and to
what extent are the dams that you are assessing, how far along
are you?
Mr. Todd. We--of course, as you said, we don't have any
responsibility for the non-Federal dams. But in reclamation, we
have 252 high and significant hazard facilities, and of those
facilities, we have assessed all of them. What we would call
our ``major mission-critical facilities,'' which are the very
top-producing power-generating dams and also very high dams, we
have used the RAM-D on. There are about 50 of those that we
used the RAM-D that was developed in conjunction with Sandia.
Those are assessed, and those are the ones that we did.
Now we have used the other ones. We have done different
priority dams and low-cost methods.
Mrs. Christensen. I yield back my time.
Mr. Lungren. I thank the gentlelady for yielding.
Mr. Dicks is recognized for 5 minutes.
Mr. Dicks. I wanted to go to the dams question. It says
here, significant information on control systems is publicly
available. It says design and maintenance documents, technical
standards for the interconnection of control systems and
standards for communication among control systems, all of which
could assist hackers in understanding the system and how to
attack them. Moreover, there are numerous former employees,
vendors, supporters, contractors and others, end users of the
same equipment, worldwide, who have inside knowledge about the
operation of the control systems.
So, Mr. Todd--and we have got information here that al-
Qa'ida has, in fact, said they are interested in the operation
of these dams. I am told--maybe you covered this earlier, but I
have got to go back to it.
We have heard the story of a hacker gaining control of some
systems of the Roosevelt Dam in Arizona, which holds 400
trillion gallons of water. What is the worst damage that could
have been done there?
Mr. Todd. In that particular situation--and that happened a
number of years ago and, of course, there have been a lot of
upgrades to that system to not allow that to happen again; that
individual did intrude, but did not get access or gain access
to any of the operation of the gates and so forth.
Mr. Dicks. Could it be done from outside?
Mr. Todd. Well, yes, there are always those possibilities
that it could be done, especially if it is hooked up to outside
systems.
We believe that is a low risk in our system because they
are not hooked up to outside systems.
Mr. Dicks. Is there encryption?
Mr. Todd. Yes, there is.
Mr. Dicks. Let us say a terrorist got control of the dam.
Is there a way to override this system at the dam?
Mr. Todd. Yes, there is. We have operators on 24 hours a
day. When we notice that the particular facilities that are
controlled are not operating in the way that we believe they
should be, we have manual controls. We do send our maintenance
people out to check those. Sometimes we take over in manual
control and operate the system manually just because there may
be a glitch or something.
So, yes, we do have ways to do that.
Mr. Dicks. Do you have a comment there at the end,
Mr. Paller?
Mr. Paller. Yes, I have a small comment. There are two
other ways to connect to these.
First of all, the word SCADA doesn't cover all the control
systems. We had a fight about that this morning. SCADA is just
the distributed system; sometimes the very big gates use other
systems called digital control systems.
I don't know to what extent those gates are not controlled
by SCADA, but controlled by digital control systems. If there
is a digital control system, most of those have dial-up access
for maintenance ports, and Bill knows a lot about this.
So this idea--SCADA is not connected, doesn't define the
whole problem. I am not saying that what--
Mr. Dicks. You are saying there are other vulnerabilities?
Mr. Paller. There are other ways of getting into those
systems besides the Internet. There are other systems that
control those gates besides SCADA systems. Sometimes they are
called DCS, sometimes they are called RTUs; they have got other
names.
Mr. Dicks. Could hackers get into those systems as well?
Mr. Paller. The FBI has reported that they already have. It
might not be true. I mean, the only data I have got is, the FBI
has reported it has.
Mr. Dicks. Interesting point.
Mr. Paller. No, listen, it wasn't--it wasn't attacked.
Mr. Dicks. Now, does the Bureau of Reclamation, do you have
control over the Army Corps of Engineers dams?
Mr. Todd. No, sir, we do not.
Mr. Dicks. So they are completely separate?
Mr. Todd. Yes, they are.
Mr. Dicks. All the private dams are separate?
Mr. Todd. Yes, they are.
Mr. Dicks. Are you working to try to develop best practices
in the industry?
Mr. Todd. Yes, we have, especially on the physical side. We
work directly with the Corps of Engineers and TVA and Homeland
Security on those systems, and we are fully engaged in that.
One of the outcomes of the Government Coordinating Council is
to work with the private side and to get information sharing
and communications going, so we believe that is working well.
Mr. Dicks. Mr. Purdy, they beat up on you pretty good
today. Let me ask you this.
We spent a couple billion dollars, several billion dollars
at the Department of Defense trying to put in place encryption
on all kinds of different defense systems.
Have you benefited from any of that? Does Homeland Security
get briefed on information from Defense about what they did to
secure their systems?
Mr. Purdy. Yes. We have a close working relationship with
the Information Assurance office within the Department of
Defense, as well as a similar entity within the National
Security Agency. So we share in the benefits of the information
that they have gleaned and share with us.
Mr. Dicks. Can you give us any examples of anything that is
been achieved?
Mr. Purdy. Well, I can't mention--I don't recall.
Mr. Dicks. If this is classified--I don't want to get into
classified information obviously.
Mr. Purdy. I can't recall specific encryption benefits, but
in those kinds of techniques, things as simple as making sure
you encrypt the data not only in transit, but at rest, and how
to protect those databases from attack are some of the examples
of things that we have learned from them.
Mr. Dicks. Any comments on this point from any of the other
witnesses?
Mr. Rush. Yes. We have actually--completely, independently,
as an industry organization, the American Gas Association got
together with a group of people and put together an open
standard. Any company can build it, and it provides a very high
level of protection, not military grade, and it is an open
standard. It is ready.
We have two manufacturers who have begun producing
prototypes. It is ready to go. We are not talking something
theoretical.
Mr. Dicks. Are people ordering it? Are companies ordering
it?
Mr. Rush. At this point they are openingSec. ordering them
in small numbers, yes, they are. But they are only ordering
them in evaluation kits, typically about five.
Until it works and people have tested it, people will be
slow to adopt them. But, yes, they are adopting them.
Mr. Dicks. Thank you, Mr. Chairman.
Mr. Lungren. I think we have about 6 minutes to get over to
the floor to vote on the first 15-minute vote.
I want to thank this panel. I think it has been very
helpful, very instructive. We make requests that all or some of
you come back at another time, because this subcommittee--I am
sure my cochair shares this--desires to continue to look at
this.
I thank you all for your valuable testimony and the members
for their questions. The members of the committee may have some
additional questions for the witnesses, and they may submit
them to you in writing. I would hope that you would answer
those in a timely fashion. The hearing record will be held open
for 10 days.
Mr. Lungren. The committee stands adjourned.
[Whereupon, at 5:40 p.m., the subcommittee was adjourned.]
A P P E N D I X
----------
Dr. K.P. Ananth Responses to Hon. Daniel E. Lungren, and Hon. Dave G.
Reichert, Letter dated November 8, 2005
I. The Threat: Probability/Impact of Attacks on SCADA Systems
1. Based on available research, how likely is an attack on a SCADA
system?
Based on a review of 120 incidents, the current likelihood of a
severe attack is low; but if the rate of incidents follows what has
been seen for the Internet in general, we forecast that the risk will
rise to a significant level in the future. Documented case histories
show that activity has increased significantly since 1988. Many of
these incidents come from the Internet by way of opportunistic viruses,
trojans, and worms, but a surprisingly large number are directed acts
of sabotage. Additionally, it is likely that there are many attacks not
being reported because many asset owners are reluctant to share or
report their experience.
SCADA systems are currently at risk from attacks stemming from a
broad spectrum of attackers ranging from common Internet threats to
directed attacks by individuals. The likelihood that SCADA systems are
attacked in a manner that results in severe consequences is dependent
on the potential attacker's motivation, intent, and expertise. SCADA
systems are vulnerable and can be exploited to result in a disruption
in service if an attacker invests enough time to learn the system
before they attack. To date, the majority of reported attacks against
SCADA systems have been the result of general Internet propagating
viruses and worms that were opportunistic in nature and not directed.
2. What cyber security failures and incidents have you seen with
SCADA networks?
Incidents to date have exposed poor security processes and
vulnerable technology implementations. The lack of general awareness as
to how the technology can be exploited has resulted in vulnerable
technology implementations and weak security practices.
In the past, incomplete security efforts and risky practices have
allowed common Internet attacks to randomly bleed into SCADA
environments. In one example, servers infected before shipping by the
manufacturer were mounted directly onto a control system network.
Security incidents impacting SCADA/control systems have been
documented in 11 sectors. The largest number of incidents has occurred
in the petroleum, power and utilities, transportation, and chemical
sectors, which combine for over 70% of the incidents observed. None of
the documented incidents have resulted in a significant event that
resulted in loss of life, major disruption of service, or economic
impacts. The US-CERT Control Systems Security Center (CSSC) has issued
a report describing the reported incidents. (US-CERT Control Systems
Security Center, Industrial Security Incidents, June 9, 2005)
3. Based on all available research, how frequently are SCADA
networks attacked?
There have been only a few reports of directed attempts to
penetrate and compromise operational control systems. However, there is
no way to know with a high degree of confidence how many attacks take
place because there is currently no formal center to report cyber
attacks on control systems. A single reporting center is operated by
the British Columbia Institute of Technology (BCIT). But reporting to
the BCIT incident reporting system is purely voluntary. The BCIT
primarily represents North America (Canada and the United States) with
several members from the UK and Australia. It is doubtful that the
reporting to the BCIT represents more that 10% of the total number of
events. The CSSC has also collected incidents from several other
reporting sources. These sources have documented approximately 120
documented cases in the past 20 years with the majority (more than 70%)
occurring in the past 5 years. Therefore, a reasonable estimate of the
number of attacks, resulting in some damage, is between 20 and 200 per
year. General cyber security monitoring at the perimeters of
organizations using power sector SCADA systems has shown a higher rate
of system probes and cyber reconnaissance activity than organizations
belonging to other sectors.
This estimate includes a wide range of possibilities because actual
incident reporting is very low. The low percentage of incidents that
get reported is due to several factors, including:
Organizations often perceive risk in reporting
security incidents
Many organizations lack the technical skill sets to
detect sophisticated intrusions or to forensically investigate
such activity
Security technology is not well-suited for SCADA
environments and existing technology have few features that
lend themselves to detect attack activity
Lack of general awareness as to the vulnerability of
SCADA systems often results in not enough attention or efforts
to detect attack activity.
The most immediate need in the arena of incident tracking is a more
effective way of reporting cyber attacks (all or at least successful)
on control systems. This enhanced reporting system needs to be a joint
effort between industry and government and needs to provide anonymity
to the reporter.
Technology trends will continue to create more vulnerabilities, and
provide greater opportunities for threat actors to access control
system networks. More interconnectivity and communication among cyber
systems will lead to increased opportunities for talented people to
breach the security systems and maliciously manipulate information or
control system functions. We also anticipate this interconnectivity and
communication capability to increase in control systems, at least for
the foreseeable future. While access to operator information and
denial-of-service attacks may cost industry money or result in
embarrassment, the manipulation of system functions using this
information can have more far-reaching consequences.
4. Is it possible to devise an attack to disable or disrupt a SCADA
network for an extended period of time? If so, what is being done to
mitigate such attacks?
Based on current testing and the knowledge of only a small number
of actual control system implementations, we believe that cyber attacks
can be devised to potentially disrupt SCADA systems (electric sector
control systems) for as long as five to seven days. However, this does
not necessarily translate into a failure of the physical system or
controlled process for the same time frame. It is possible for a
sophisticated attack to poison databases and files over time that would
require a system re-build and re-configuration before the control
system would function normally. More research is needed to investigate
if cyber attacks can cause significant failures in long lead time
physical equipment, such as transformers and generators. Similar
studies are also needed in other sectors such as water, transportation,
and chemical plants to assess equipment impact and downtime.
Our cyber security researchers have demonstrated the ability to
physically destroy many of the IT components used in the control of a
SCADA system. The practice is commonly referred to by hackers as
``bricking'' a box. There are many ways to require that a SCADA system
be rebuilt from the ground up. Additionally, if the attacker plants a
program in the backup sets ahead of time, the system will just destroy
itself again as soon as it is brought back online. The attacker can
also plant programs in non-essential equipment such as card readers,
and printers that are unlikely to be found. The result is long-term
disruption of service.
Many of the physical devices are set to automatically shut off at
preprogrammed points to protect the devices from overheating/
overdriving/overworking. In some instances an attacker can reset those
points and drive the hardware to failure. Rhythmically turning on and
off a 480-volt motor can destroy it. Operating a valve hundreds of
times a second can destroy it. Flow-cooled pumps will overheat and fail
if the valve is closed while the pump is running. Many other scenarios
are easy to find and exploit.
Based on our testing in a representative configuration (an electric
sector EMS system) established in the test beds, it is possible to
disrupt system operation through cyber attack. The duration of the
disruption will depend to a large extent on the types of attacks
executed, the specific owner/user's system configuration, backup
capability, and response/recovery practices. Mitigation efforts to date
have focused on identifying specific vulnerabilities by examining
representative systems in the test beds and providing information to
system vendors who then eliminate the vulnerabilities in their
products. Work in the test beds is also helping to identify the best
practices that can be implemented by both the vendors and the users in
making their systems less vulnerable. A significant effort is being
made to enhance owner and vendor awareness of the methods for reducing
vulnerabilities.
5. (Not assigned)
6. Electric power is important for nearly all the things that Americans
do--from businesses to schools to government to many forms of
recreation. Has your research shown that the SCADA systems that control
our power generation and distribution are fully protected from attacks
launched from the Internet? If not, what kind of damage do your
researchers believe smart, well researched attacks could cause?
Although some SCADA systems that control power generation and
transmission currently have some form of cyber protection, power sector
SCADA systems are not ``fully protected'' from Internet-launched
attacks. Research has shown that the majority of vendor solutions are
vulnerable to a cyber-based attack coming from the Internet and through
the surrounding corporate network that could result in a complete loss
of system control. Those attacks were successfully demonstrated despite
the use of common configuration practices and the use of available
security technologies (IDS, Firewalls, etc). For obvious reasons the
majority of this research has not been replicated in the field but INL
has the ability to create very large scale control system and physical
infrastructure simulations in both the electric and chemical processing
sectors.
We have also seen evidence of SCADA systems being vulnerable to
non-expert-based attacks. In fact, non-directed common and
opportunistic threats, such as viruses and worms, have impacted SCADA
systems. Considering a random threat such as a virus can impact a SCADA
system, a well resourced and motivated threat actor could compromise a
control network and cause significant disruption to power SCADA
systems. The disruptions may or may not result in wide-spread power
outages depending on how much the attacker learned once inside of the
target's control system. Certainly, a directed attack can result in
injected commands being passed through the SCADA system to breakers in
the field possibly resulting in breakers taking lines out of service.
Assessments performed in the test beds show that typical control
systems can be compromised from the Internet if the attacker has some
understanding of the system. Much of that system information can be
obtained by a patient study of open source information. A well-
orchestrated attack could provide the attacker with the capability to
take over the operator's function, potentially without the knowledge of
the operator. While strongly influenced by system configuration and
operating policies, there is the potential to cause damage to equipment
through the manipulation of operating and safety limit set points.
7. (Not Assigned)
8. We've heard a lot about the impact of a terrorist attack on a
control system. But as we saw during Katrina, natural disasters can
cause devastating impacts to our control systems infrastructures too.
What kind of impact would natural disaster have on control systems in
California (earthquakes), Oregon (Tidal waves/Tsunamis), The Gulf Coast
(Hurricanes), elsewhere?
Any event, whether manmade or natural, resulting in the destruction
of physical equipment and the loss of supporting services like water,
power, and communications can negatively impact SCADA systems.
Anecdotal information and data emerging from hurricanes Katrina and
Rita are showing that, for SCADA and other control systems (and other
utility operations), the need to plan and prepare for an ``all hazards
approach,'' rather than more narrowly defined scenarios, is crucial.
We learned from Hurricane Katrina that the main impact to a control
system from a natural disaster is the remote entities that the system
connects with (e.g. customers, substations, transmission lines). After
the August 29th landfall of Hurricane Katrina in Louisiana, 2.7 million
customers were without power, 263 substations and 181 lines were not
operating. As of September 22nd, less than 250,000 customers are
without power and 19 substations and 25 lines remain out (data from the
Office of Electricity Delivery and Energy Reliability U.S. Department
of Energy, Hurricane Katrina Situation Report #42, September 23, 2005).
The control centers themselves are normally less vulnerable than the
remote devices that are being controlled and queried for status.
The ability for a control system to minimize impact from a natural
disaster is directly related to the system owner's continuity of
operations, disaster recovery planning, and overall preparedness to
handle natural disasters as discussed in the US CERT website (US-CERT
Informational Paper September 16, 2005, produced by the US CERT Control
Systems Security Center, Hurricane Katrina Control System Assistance
http://www.us-cert.reading_room/KatrinaCSA.pdf).
The control system is only as good as the data it can receive. With
limited view and communications, the systems? components and the
applications designed for automatic control cannot be used properly
without subject mater experts making the decisions. In the case of
Katrina, the restoration process was hampered by the other
communications outages of telephone and wireless.
The National Infrastructure Simulation and Analysis Center (NISAC)
provides advanced modeling and simulation capabilities for the analysis
of critical infrastructures, their interdependencies, vulnerabilities,
and complexities. It would be helpful to study lessons learned during
Katrina on the effectiveness of the NISAC models.
II. The Public/Private Relationship in Developing a SCADA Solution
1. I understand the National Labs are conducting extensive research
into SCADA and Control Systems. What resources are you currently
lacking? How are you coordinating these efforts with the private
sector? What can the federal government do to provide more resources?
Needed Resources: INL recommends a 5-year funding profile that
allows the development of long-term programs to support critical
infrastructure sectors immediate and long-term complex SCADA
challenges. The uncertainty of year-to-year funding and funding delays
at the beginning of the fiscal year negatively impact our ability to
provide sustained research to identify vulnerabilities and to develop
solutions to fix vulnerabilities aligned with asset owner and vendor-
driven timelines.
Sustained funding will allow us to successfully decrease risks to
control systems by conducting ongoing tests to identify vulnerabilities
and develop mitigations, raising awareness and helping organizations
develop the right mind set to protect SCADA systems, gaining access to
more credible incident information, conducting in depth research and
testing to explore possible consequences and outcomes, and monitoring
the cyber underground to gauge their knowledge of and interest in SCADA
systems.
Private sector coordination efforts: INL is working directly with
asset owners and vendors to evaluate their system vulnerabilities and
implementing mitigation steps. These evaluations are protected using a
nondisclosure basis.
INL is engaging national experts from industry, national labs, and
academia in dialog to keep current on allied research and best
practices and to share that knowledge with industry. In FY-05, we
conducted nine regional workshops and participated in the Process
Control Forum. These interactions directly impacted 280 asset owners.
Our industry outreach program includes training and awareness
demonstrations of the means and effects of a cyber attack on control
systems. These demonstrations and training activities are ongoing with
positive feedback from industry and government participants. These
include live demonstrations of attacks/effects on small scale
representative control systems for chemical and electric system
processes and cyber security--control systems training uses these tools
and subject matter experts.
Additional federal government resources: Along with sustained 5-
year funding, designate INL as a National Center of Excellence and User
Center for SCADA, Cyber Security, and Critical Infrastructure
Protection. The Center would be modeled after existing National User
Facilities at other DOE National Labs, such as the High Temperature
Materials Laboratory at Oak Ridge National Lab or the Light Source
Facility at the Brookhaven National Laboratory). The Center designation
would capitalize on INL SCADA test beds and full scale infrastructure
assets, build on our proven track record with asset owners and vendors
to identify and mitigate cyber vulnerabilities, and provide an
independent, scientific organization that tests and validates the
vulnerabilities and identifies solutions. The result is federal/private
partnerships with high value to the critical infrastructure owners and
their vendors.
With long-term dedicated funding, INL can move from the current
research approach, which focuses testing on specific attacks as a
method of raising vendor awareness, to conducting extensive assessments
in a comprehensive fashion. We would develop consistent methodologies
and system rating approaches that would apply across all vendors and
develop quantitative measures to verify the return on investment of
research dollars that directly impact industry and taxpayers. To that
end we would devote research focus to develop a realistic threat
assessment methodology and then apply it to create an open, industry-
acknowledged threat model for contingency planning.
2. (Not assigned)
3. It has been widely reported that both industry and the federal
government find it difficult to estimate the economic impact of a cyber
security attack. Has the lack of actual quantifiable damages made the
private sector leery of investing in cyber security?
There has long been widespread agreement that the published
estimates of cyber-attack costs have little credibility. In April 2004,
the Congressional Research Service Report on The Economic Impact of
Cyber-Attacks concluded ``No one in the field is satisfied with our
present ability to measure the costs and probabilities of cyber-
attacks.'' But the report resulted in limited research to address the
measurement need. The research programs most directly addressing the
need for better assessments of cyber-attack consequences are the
programs of the U.S. Cyber Consequences Unit, a small independent
agency established by the DHS in August 2004. The first of the larger
US-CCU reports will be available for limited circulation release in
early February 2006.
The lack of economic consequence data and security metrics has led
to a variety of concerns about the possibility of a successful attack
and its associated economic impact. Currently, there is no consensus
about the level of resources that should be devoted to control systems
cyber security. Standards and associated business cases are being
developed that will help industry better evaluate the risk to their
systems. Even with this lack of documented cases of quantifiable
damage, attacks occur. For example, recent malware attack (Zotob) on
multiple sites of a large manufacturing company resulted in loss of
production time.
These types of attacks increase asset owners' awareness that they
too could be the target of a potentially crippling attack; thus,
investments are being made in the private sector. These investments
tend to be dependent on the extent of awareness of cyber intrusions and
the liability posed by denied services or business losses faced by
individual companies as well as customer impact. Critical
infrastructure sectors, such as electric utilities, chemical companies,
oil and gas companies, and banks and financial institutions, realize
the potential impact of cyber threats but the investments and attention
paid is not uniform across the sectors. Cyber security concerns
resulting from easy electronic access to accounts in the Banking and
Financial Sector are addressed USA Today's November 2, 2005 first page
article, ``Cyber crooks break into online accounts with ease''. In the
Electric Sector, the required connectivity with neighboring systems
creates a weakest link problem for the overall network of
interconnected SCADA systems. The larger or more progressive utilities
will suffer from weaknesses presented by smaller, resource-constrained
neighbors.
Several industry associations, such as the Chemical Information
Data Exchange (CiDX), the Water Environment Research Foundation (WERF),
and the American Association of Railroads, are promoting cyber security
among their subscribers. The Department of Homeland Security Control
Systems Security Center (CSSC) established an Industry Interest Group
to discuss asset owner's perspective of cyber security. Members of this
group reported that at the operations levels within their company's
organization, cyber security is important. However, at the board of
director's level, cyber security seems less important because they may
not see any risk to bottom-line profits. The group also reported that
awareness communication tools would be helpful in convincing their
management to invest in SCADA security, even though the perceived risk
may be low at this time.
The reason the National Cyber Security Division of the DHS
established the US-CCU, with the support of the National Communications
System and help from the DHS Private Sector Office, was that both
corporate executives and government officials regularly reported they
could not justify larger cyber security budgets without better
information on the likelihood and costs of possible cyber-attacks.
4. (Not Assigned)
5. Can you tell us specifically how your research on SCADA has, to
date, impacted the way SCADA systems in the field are secured, and what
percentage of those systems have been impacted? If that's not a big
number, what is stopping us from putting the results of your research
into practice in the field?
A result of our assessment work in the test beds is the
identification of best practices that can be used to mitigate
vulnerabilities by taking advantage of the capabilities already
existing in the SCADA systems. Examples include ensuring fully patched
operating systems, improving password management practices, and
implementing layered security defenses (firewalls, DMZs).
SCADA system vulnerabilities identified through assessments
performed in the test beds have been communicated to the manufactures
and users of those systems. In all cases, the vendors have taken quick
action to incorporate system modifications to mitigate the identified
vulnerabilities in their new systems, but only 5% of installed systems
are new systems. Thus implementing enhancements in currently installed
systems requires that owners be made aware of the vulnerabilities
within their systems and the mitigating methods that are available to
them.
More than 230 user representatives from over 100 major electrical
industry owners/users of SCADA systems have been made aware of typical
vulnerabilities and methods for security enhancement. The percentage of
the industry that is represented by 100 owners is difficult to answer,
but in very general terms we can say that they control approximately
80% of the power on the grid. This communication has been achieved
through presentations and discussions in numerous electrical industry
user group meetings and conferences. In addition to electrical industry
interactions, workshops, demonstrations, training, and presentations
have been provided to audiences responsible for control systems used
across the Nation's critical infrastructure. In aggregate, these
various forums have been attended by more than 7500 people from vendor
and user companies.
In addition to assessments, cyber security awareness workshops in
nine regions involved 480 industry participants during FY-05 have made
the industry more cognizant of the need to strengthen their SCADA
systems. In FY-06, we will be providing asset owners additional tools
to strengthen SCADA security through vulnerability assessments both in
test beds and at participant selected facility locations. The value of
the INL work, as perceived by a sample of industry/end users, has been
previously stated (see INL's written testimony of October 18, 2005, to
the same Subcommittees).
We do not have access to data that would quantify the extent to
which system owners are implementing our recommendations into their
administrative and hardware/software management policies. This is
typically information that is held close by the asset owners for
competitive advantage reasons. Because the deployment of new systems
occurs rather slowly (estimated at 5% annually for the installed
infrastructure) the users, working with their vendors, can also design
and implement mitigations specific to their systems. Thus the
information we provide can be used to upgrade and improve configuration
and management of currently installed systems.
The reason for relatively slow system upgrades is the high cost and
the lack of a strong business case (bottom line dollar impact) to
justify both the expenditure for improvements and to justify requests
for recovery through the rate base. A frequently raised issue is that
if the requirements for security upgrades were mandated through
regulation, the asset owners would have a stronger basis for requesting
rate relief. However this brings with it the added burden of additional
regulation to the industry and is therefore not strongly supported by
industry.
6. What has the money we have already spent on SCADA research done
to improve SCADA security in the field?
The work performed and supported by the Department of Energy
National SCADA Test Bed (NSTB) in the Energy Sector and the Department
of Homeland Security Control Systems Security Center (CSSC) Program on
the other sectors, have improved security at critical infrastructure
facility sites in significant ways:
Awareness: As a part of the mission for both the NSTB
and the CSSC, cyber security awareness has increased in
industry and government. Information on potential threats,
vulnerabilities, and mitigation of cyber attacks on control
systems has been disseminated through workshops, outreach, and
training events at conferences, user groups, and invited
sessions. The increase in awareness of the potential for real
and serious impact to facility operations have resulted in
asset owners performing reassessments of their cyber security
for control systems.
Assessment and Testing: CSSC and NSTB are engaged in
performing assessments of major control system SCADA vendors?
current products to identify both vulnerabilities and
mitigation. Some of the vendors have taken steps to eliminate
the identified vulnerabilities and shared the information with
their users. Working closely with the vendors and the user
community, the CSSC and NSTB provide a path to rapidly identify
and facilitate the use of this information to increase the
protection from cyber attacks. The success of these
relationships act as models to both the vendor and user
communities to work with these DOE and DHS programs. Several
site specific assessments have also been conducted at the
request of asset owners. Results of these assessments provide
direct and specific input to increasing SCADA security at those
sites.
Technology Development: A key element of the CSSC
program is the identification and quantification of risk that
supports a business case to the asset owner for the policy,
time, and equipment investments to reduce risk to acceptable
levels. The characterization of vulnerabilities (control and
network systems), consequences (safety and national security),
and threats (beginner level to hostile nation state) coupled
with the cost of implementation of safeguards is a necessary
step in developing risk models and the business case. The CSSC
is active in working and coordinating efforts with industries,
industry and trade associations, government agencies, and
academia to identify gaps in technologies and standards to
apply to both current and legacy critical infrastructure
control systems. While these efforts are emerging, the broad
exposure of this work and participation of the stakeholders
will produce improvements in SCADA security that meet the need
for information protection coupled with business constraints
and will increase security awareness.
US-CERT Support: The United States Computer Emergency
Response Team (US-CERT) provides response and capabilities to
support government and the private sector dealing with cyber
threat and attacks to the Nation's network communications and
computing infrastructure. The CSSC augments this capability by
providing expertise in control systems and the potential
vulnerabilities and impacts of cyber attacks. The CSSC has a
broad reach of assets within the national laboratories and
private sector to assess situational awareness during specific
response to events reported to the US-CERT. The CSSC, as a part
of the US-CERT in these activities, can issue alerts to be
distributed at a national level given that their may be real
and significant threats to control systems for certain sectors
or user communities. The goal of this capability is to provide
another level of information to those asset owners to increase
SCADA security to threats.
7. Is there any risk of duplicating efforts with the lab beds at
Sandia and Idaho and other research around the country?
INL is directly involved in two programs, the National SCADA Test
Bed sponsored by DOE/OE for the Energy Sector and the CSSC Program
sponsored by DHS/NCSD for the other sectors. We are working with Sandia
and others to complement what is needed to carry out the objectives of
these programs and there is no duplication of efforts. Also to prevent
the duplication of efforts, the sponsors (DOE/OE and DHS/NCSD) review
the scope of work on the NSTB and Control Systems Security Center
Programs.
INL and Sandia each have unique and complementary SCADA
capabilities. INL focuses on evaluating Cyber Security vulnerabilities
of SCADA systems deployed in operational facilities and validating
solutions; and penetration testing of control systems. Also INL has on-
site, full scale infrastructure systems such as electric transmission
systems, substations, a pilot chemical plant and communications test
beds that enable field scale evaluations. Sandia, on the other hand,
has information technology red teaming and assessment capability,
cryptography, and bench scale testing capability complementing INL's
capabilities. The two Labs recognize their strengths and collaborate to
provide the service needed to support asset owners and vendors.
Because of the number and diversity of infrastructure facilities in
the US requiring SCADA/Cyber security and the level of coordination of
efforts between INL and Sandia, there is great value in having two
national labs with capability and capacity to provide a wide range of
assessment services to asset owners.
INL, as the lead lab for the control system cyber security program
coordinates efforts between labs utilizing specific expertise,
facilities, and capabilities at each laboratory to perform its work. In
January of 2005, a Leadership Steering Group was organized and consists
of members from Idaho National Lab (INL), Sandia National Lab (SNL),
Pacific Northwest National Lab PNNL), and Lawrence-Livermore National
Lab (LLNL). The Group meets on a quarterly basis to discuss the
direction of the program, coordinate efforts and deliverables, and
identify expertise that is needed to solve issues and challenges. Ideas
are exchanged and security products developed for various governmental
customers are shared.
III. The Federal Government's Role in Cyber Security
1. (Not Assigned)
2. (Not Assigned)
3. (Not Assigned)
4. (Not Assigned)
5. There are several SCADA test beds across the country. Is there
any risk of duplicating efforts with the lab beds at Sandia and Idaho
and other research? Is there anyway to consolidate these efforts?
INL is directly involved in two programs, the National SCADA Test
Bed sponsored by DOE/OE for the Energy Sector and the CSSC Program
sponsored by DHS/NCSD for the other sectors. We are working with Sandia
and others to complement what is needed to carry out the objectives of
these programs and there is no duplication of efforts. Also to prevent
the duplication of efforts, the sponsors (DOE/OE and DHS/NCSD) review
the scope of work on the NSTB and Control Systems Security Center
Programs.
INL and Sandia each have unique and complementary SCADA
capabilities. INL focuses on evaluating Cyber Security vulnerabilities
of SCADA systems deployed in operational facilities and validating
solutions; and penetration testing of control systems. Also INL has on-
site, full scale infrastructure systems such as electric transmission
systems, substations, a pilot chemical plant and communications test
beds that enable field scale evaluations. Sandia, on the other hand,
has information technology red teaming and assessment capability,
cryptography, and bench scale testing capability complementing INL's
capabilities. The two Labs recognize their strengths and collaborate to
provide the service needed to support asset owners and vendors.
Because of the number and diversity of infrastructure facilities in
the US requiring SCADA/Cyber security and the level of coordination of
efforts between INL and Sandia, there is great value in having two
national labs with capability and capacity to provide a wide range of
assessment services to asset owners.
INL, as the lead lab for the control system cyber security program
coordinates efforts between labs utilizing specific expertise,
facilities, and capabilities at each laboratory to perform its work. In
January of 2005, a Leadership Steering Group was organized and consists
of members from Idaho National Lab (INL), Sandia National Lab (SNL),
Pacific Northwest National Lab PNNL), and Lawrence-Livermore National
Lab (LLNL). The Group meets on a quarterly basis to discuss the
direction of the program, coordinate efforts and deliverables, and
identify expertise that is needed to solve issues and challenges. Ideas
are exchanged and security products developed for various governmental
customers are shared.
6. (Not Assigned)
IV. The Federal Role in the Future
1. Based on your knowledge of the SCADA research field, what are
the most promising technological breakthroughs you see that can protect
our SCADA systems in the short term? I realize there are no silver
bullets, but please list the solutions that will actually work to
protect our SCADA systems.
Various emerging technologies show promise in protecting control
systems. Deep packet inspection engines (optimized to detect control
system packets) can guard for commands or injects traveling through
unauthorized avenues like the organization's perimeter or corporate
network. Memory cache integrity technologies can be used to detect
malicious events like buffer overflows. Secure authentication
approaches applied to SCADA protocols and emerging low-overhead
encryption techniques are also promising. The optimization and use of
these emerging security technologies should reduce some of the risk
SCADA systems now face. In order to bring these technologies to bear
more testing environments need to be used to test general IT security
solutions and enhance them to work in control system environments.
Near-term security enhancements can be most effectively implemented
through taking advantage of existing technologies. This can be done
through the definition and implementation of security policies based on
the best practices identified in the test bed efforts and in industry.
Best practices include defining the electronic perimeter, setting up
layered defenses, monitoring communication traffic for anomalies (such
as with intrusion detection and prevention devices), and establishing
strong password management and system patching policies. Encryption
technologies should be applied to eliminate plain text communication
that can be monitored by an intruder to obtain system knowledge.
On a longer term basis, secure programming techniques should be
used in application code development as is now being done for operating
systems and embedded applications.
Much knowledge exists, but there is a gap between general IT
security and SCADA security. SCADA systems have to be ultra-reliable
and ultra-stable. If cyber-security is going to take hold in SCADA
networks, the following must take place: (1) a testing location where a
utility can test their configurations with expert support and advice
must be developed and (2) a user community where users of the same
SCADA system with the same problems can critique their architectures
and perform peer reviews must evolve.
2. How do we make rapid progress in improving security in the
field?
Increasing awareness among asset owners and vendors should be a
priority because vendors must eventually implement the security
measures. Another priority should be providing the ability to test the
systems in an impartial manner. Third is providing the tools that are
needed to mitigate the vulnerabilities and secure the systems. Finally,
some consideration is required for financial incentives to accelerate
cyber security implementation by asset owners. In all of these steps we
should also look at the interlinked aspects of information technology,
control systems and telecommunication and take a systems approach to
dealing with this challenge. The key to success lies with increasing
industry awareness, and industry associations can play a critical role.
Many of these groups have already seen the need for improving cyber
security in control systems and have started working groups or sub-
committees to address the issues and share information with their
subscribers. As NCSD shares vulnerability findings and provides best-
practices for mitigation to these associations, they are transmitted to
their members and mitigations are implemented.
A good example of the security initiative within industry is the
Chemical Information Data Exchange (CiDX). In January 2003, CiDX
started the Chemical Sector Cyber Security Program. This program has a
sub-committee that is devoted specifically to cyber security for
control systems. They recently recommended that the CiDX subscribing
companies perform self-assessments of their control system security
posture. Several companies reported results at the October CiDX General
Membership meeting in Houston. While, these self-assessments are still
immature, their willingness to improve their security posture is
commendable. The NCSD has developed a self-assessment tool to help
associations like CiDX improve the effectiveness of their self-
assessment process. The tool will assist asset owners to focus on the
critical cyber security requirements and associated compliance
strategies to achieve improvements in security. In FY-06, the self-
assessment tool will be piloted with several asset owners in multiple
sectors. After the piloting effort, NCSD will improve the tool, provide
training at workshops in the various associations, and commence wide-
spread distribution and use of the self-assessment tool. This will give
asset owners specific measures for immediate implementation and
reduction cyber security risk.
Rapid progress is based upon a multi-tiered approached that
involves diverse stakeholders. This includes system integrators,
vendors, and asset owners. Increasing security in the field will
require each one of these stakeholders to develop better integration
requirements that include improved security, hardened vendor systems,
and increased situational awareness, respectively. Asset owners need to
increase their awareness to control system cyber security and the
inherent reliability benefits to addressing security, thereby
requesting that secure system be purchased and integrated into the
field.
3. (a) Has the federal government advocated for standards
establishing a minimum floor for securing control systems?
While the argument could be made for a minimum floor standard, this
may not be the solution for the long term. Since 85% of our critical
infrastructure is owned by the private sector, it is their
responsibility to adequately protect their assets and deliver the
services and products to the customer at large. The liability that
could result from a federally mandated minimum standard argues against
such a standard. Also, the need for continuous improvement is
disincentivized by a minimum standard. In our view, industry groups
working together should come up with the best practice for their
industry segments. The electric utility, chemical industry, and oil and
gas industry have all come up with some type of best practice and they
should be encouraged to make more widespread use of these practices.
Similarly, other industries should come up with best practices for
their segments with help from the federal government in terms of
testing vulnerabilities and developing mitigation measures.
The DHS (CSSC Program) and DOE (NSTB Program) both include tasks to
support improvements to industry security standards. In addition to an
ongoing review of standards applicable to control system security (with
the goal of identifying areas that should be strengthened), activities
include support to drafting ISA's SP-99 and a technical review and
assessment of the standard for Secure ICCP.
3. (b) What would a minimum floor look like?
A minimum baseline standard should address areas that are important
to cyber security in general, with an additional emphasis on areas that
are of particular concern to control system security. Control systems
are complicated and varied depending on their application. Developing
standards that address security needs has begun (as outlined below,
Question 3-e) but addressing the hundreds of needs for securing the
complexities of control systems will require a large concentrated
effort.
Topics that should be addressed include: the assessment of risk,
development of a security policy, organization of information security,
management of assets, human resources, physical and environmental
security, management of operations, access control, the acquisition,
development, and maintenance of process and information systems,
incident and business continuity management, compliance with legal and
company policies. Standards should also address next generation systems
to help ensure that security in ``built into'' emerging components and
systems.
Another area of concern is system integrators. Standards must also
address network architecture to ensure that security vulnerabilities
are eliminated at the system level.
The CSSC has developed a cyber security protection framework that
includes hundreds of high-level security requirements for the various
components and communication links in control systems. These
requirements have been compared with the myriad of existing cyber
standards to identify gaps and overlaps in these standards. In FY-06,
the findings of this review, along with continued reviews, will be used
to recommend specific changes and improvements to the various standards
bodies.
3. (c) Have industries leaders begun the process of developing
those standards already?
Several industries, particularly chemical, oil and gas, and
electrical, have made great strides in the development of control
system cyber security standards. In addition, professional
organizations and government bodies have contributed to the development
of these standards.
3. (d) Has the government established any ``best practices'' that
can be modeled by industry?
As mentioned above (Question 3-b), CSSC has collected an initial
set of industry best-practices for complying with security requirements
and standards. NSTB program is developing best practices aimed at
mitigating the common vulnerabilities discovered during control system
testing.
Through both NSTB and CSSC Programs, best practices are being
identified and shared with industry as stated in II-5.
3. (e) What other standards activities are being developed besides
AGA 12?
Several cyber security standards aimed at industrial control
systems have been developed or are in the process of development. Some
of these may not be considered as standards in the strictest sense, but
still provide guidance and direction. These include:
AGA 12--The American Gas Association is in the process of
developing a series of four standards recommending practices designed
to protect SCADA communications against cyber attacks. To date, Parts 1
and 2, which address Cryptographic Protection of SCADA Communications,
are still in draft form.
API 1164--The American Petroleum Institute released this standard
on SCADA security to provide guidance to the operators of oil and gas
liquid pipeline systems for managing SCADA system integrity and
security. This document was released in September 2004.
CIDX--The Chemical Industry Data Exchange has developed a Guidance
for Addressing Cybersecurity in the Chemical Sector Version 2.1. This
document describes key elements of a cybersecurity management system in
the chemical sector.
IEC 62351--The International Electrotechnical Commission is in the
process of developing ``Data and Communication Security.''
ISA TR 99 Parts 1 and 2--The Instrumentation, Systems and
Automation Society (ISA) has published two technical reports addressing
control system security with suggestions for securing control systems
against cyber attack.
ISA SP99 Parts 1 and 2--ISA is in the process of developing two
control system cyber security standards. These standards, still in
draft form, will provide requirements for securing control systems.
NIST SPP-ICS--NIST has developed and released a System Protection
Profile (SPP) to formally state security requirements associated with
industrial control systems (ICS).
NIST 800-82--NIST has developed SP800-82, a Guide for SCADA and ICS
Security. It is in draft form with scheduled release January 2006.
NERC 1200--The North American Electric Reliability Council (NERC)
has developed and released this temporary standard to establish a set
of defined security requirements related to the energy industry and to
reduce risks to the reliability of the bulk electric systems from any
compromise of critical cyber assets.
NERC CIP-002 through--009--NERC is in the process of developing a
series of standards aimed at entities performing various electric
system functions. When released, it will replace NERC 1200.
4. (Not Assigned)
5. (Not Assigned)
6. (Not Assigned)
7. Some have mentioned the value of a ``vendor'' incentives system
that would provide tax and other financial incentives to manufacturers
who are producing control systems that are already in ``best
practices'' compliance. How feasible is this, and have there been
evaluations of the cost to the federal government?
The first step in incentivization is enabling full reporting and
disclosure of cyber security incidents, without attribution, similar to
the FAA's Airline Pilot Reporting System. Included in this Cyber
Security Reporting should be disclosure of the stringency level and
thoroughness level of the reporting and assessments, so the frequency
and magnitude of the problems can be analyzed. Then appropriate
mitigation steps and incentives for implementation of these steps could
be developed. With this incident information, other incentive options
could be considered in light of the overall risk/benefit ratio.
Another incentive would be to enable independent third-party
testing and evaluation of control systems and techniques to mitigate
vulnerabilities as is now provided through the DOE/NSTB Program to
utilities and through the DHS/CSSC Program to all other industry
sectors.
The feasibility and cost of incentives would need to be studied
closely to ensure the approach provided the right reward to maximize
responsible action by vendors. The best vehicle, approach and resulting
cost to implement have not been studied.
8. (Not Assigned)
V. Dam Security
(None Assigned)
Donald Andy'' Purdy Responses to the Honorable Bennie G. Thompson
Questions
THE THREAT: PROBABILITY/IMPACT OF ATTACKS ON SCADA SYSTEMS
Based on all available research, how likely is an attack
on a SCADA system?
Response: Attacks are already occurring against Supervisory Control
and Data Acquisition (SCADA) systems/control systems; however, the
number of incidents reported is few and the consequences associated
with these reported attacks are generally not very significant. The
NCSD Control System Security Program (CSSP) has reviewed data on
approximately 120 documented cyber incidents against SCADA/control
systems over the last 20 years. This data shows that the number of
cyber attacks reported against SCADA/control systems has been
increasing over the last several years and also shows that a larger
percentage of attacks are coming from external sources as opposed to
internal sources.
As SCADA/control systems have greater interconnectivity to
information technology (IT) systems external to the SCADA/control
systems operating environment and increase their utilization of common
open standards and protocols, the exposure of systems to outside
entities and the number of vulnerabilities present in the control
system environment will continue to increase.
Insufficient data currently exists to accurately calculate the
likelihood of a successful cyber attack against a SCADA/control system
that would result in a catastrophic consequence. However, based on
current scenarios developed by industry and the National Labs, the
National Cyber Security Division (NCSD) believes that as the number of
vulnerabilities, the number of people with intent to cause the U.S.
harm, and the number of people with sufficient skills and capability to
successfully execute an attack continue to increase, the likelihood of
a successful cyber attack of significant consequence against SCADA/
control systems will continue to rise. The NCSD CSSP is working under
the assumption that a cyber attack resulting in a significant
consequence is likely to occur some time in the future. We are
aggressively pursuing mitigation remedies to reduce the likelihood of
cyber attacks on SCADA/control systems.
NCSD is establishing a control system cyber attack response center
through the United States Computer Emergency Readiness Team (US-CERT)
with technical response teams active within the CSSP. The Cyber Storm
exercise beginning in February 2006 will provide additional information
on readiness and response capabilities and needs.
NCSD is also working with the Intelligence Community to better
collaborate on SCADA/control systems threat requirements and provide
input on intelligence products.
Based on all available research, how frequently are SCADA
networks attacked?
Response: Historically, there has been no consensus on a formal
center in the U.S. for all critical infrastructure owners and operators
to report cyber attacks against SCADA/control systems. US-CERT recently
initiated efforts to serve as the central focal point for the nation's
critical infrastructures to report SCADA/control systems cyber
incidents and vulnerabilities.
A reporting center operated by the British Columbia Institute of
Technology (BCIT) also accepts voluntary submissions of SCADA/control
system incidents. Owners and operators of U.S. critical infrastructures
are hesitant to report SCADA/control system cyber incidents both
because of concerns about how the information could potentially be used
to harm the reporting organization, and also due to the absence of a
clearly designated place to report cyber incidents.
The NCSD CSSP combined cyber incident information from BCIT with
information from other sources to examine approximately 120 documented
cases occurring over the past 20 years. A majority of these reported
SCADA/control system incidents (>70%) have occurred in the past 5
years. However, it is widely viewed that the number of incidents are
highly underreported. We are working with SCADA/control system vendors,
owners and operators to raise awareness and increase cyber incident
reporting to the US-CERT.
I am interested in your assessment of the type of damage
that you believe can actually result from a terrorist attack on SCADA
systems. I think many people were shocked when on September 11, 2001,
they learned that a single airplane could cause one of the World Trade
Towers to collapse with huge loss of life. What are the corresponding
scenarios for catastrophic damage that can be caused by someone who has
taken the time to learn to control SCADA systems?
Response: Intermittent or properly timed loss of control of a
critical infrastructure control system can enhance the probability of
incorrect operator responses, which can lead to accidents with serious
physical results, such as fire, explosion, collisions, or loss of
production.
Two historic events affecting critical infrastructures where
control systems could have played a contributing role include
explosions at the Piper Alpha North Sea Platform and the Texas City oil
refinery. The Piper Alpha platform explosion in July 1988 killed 167
and resulted in losses which are estimated up to $15.2 Billion US.
Although there was a combination of events that lead to this accident,
incorrectly interpreted signals and early loss of the control room
contributed to the disaster. The March 23, 2005 Texas City oil refinery
explosion killed 15 and injured 170, and cost close to $1 Billion US.
This accident did not involve a cyber attack, but the accident evolved
as a result of the misinterpretation of signals and indicators, which
could be affected by a cyber attack.
The following are some examples of scenarios that show how cyber
intrusions could result in physical damage, loss of life, environmental
damage, economic loss, and/or loss of production in our nation's
critical infrastructures.
--The breach of security controls in the transmission mechanism for
a regional power grid system could potentially allow a strategic attack
to develop into a widespread blackout due to the unique cascading
aspects of power transmission. Although the August 2003 East Coast
blackout was not caused by a cyber attack, the failure mechanisms that
caused that blackout are similar to those that could be achieved
through a cyber attack.
--The readings on chemical mixing tanks during the batch process
could be tampered with by unauthorized network intrusion, forcing
lethal and highly combustible reactions to occur without warning to the
operators. Misinformation, exacerbated by improper response, is the
cause of many industrial accidents.
--Rogue access into the railway switching system within a major
city could cause significant gridlock to commuter traffic and import/
export functions or potentially result in a train collision.
--In a blended physical and cyber attack, quality and safety
triggers in a metropolitan water facility could be subtly compromised
allowing for normally unallowable levels of toxins or chlorine to be
distributed into the city reservoirs and pumping systems.
--According to the Network Reliability and Interoperability Council
(NRIC),\1\ the growing use of Voice Over Internet Protocol (VOIP) and
the interconnected nature of networks pose an increasing risk to the
telecommunications infrastructure, in part because internet-based
protocols are not as robust against security breaches as is traditional
telephone technology. If operations centers or network management
functions are compromised by combinations of cyber and physical attacks
there could be a cascading effect that disrupts the communications
capabilities of consumers, businesses and emergency first responders.
---------------------------------------------------------------------------
\1\ The Network Reliability and Interoperability Council (NRIC) is
a partnership of private sector entities and the Federal Communications
Commission (FCC) that develops recommendations designed, in part, to
assure optimal reliability, security and sustainability of the nation's
telecommunications infrastructure during periods of exceptional stress,
including terrorist attacks or similar occurrences. http://
www.nric.org/
THE FEDERAL GOVERNMENT'S ROLE IN CYBER SECURITY
We saw during Hurricane Katrina that the federal
government is unprepared to respond to a large natural disaster. Today
we've heard about the devastation that may be caused if a terrorist or
a natural disaster hits our control systems. Just last week, a headline
in the New York Times read: ``US cyber security due for FEMA-like
calamity?'' Are we prepared for a cyber attack on our control systems?
Similarly, if a natural disaster hits our control systems, are we
prepared to respond to that?
Response: The NCSD CSSP is being proactive in preparing for events,
both natural and man-made, that could potentially disrupt our nation's
control systems and the critical processes and functions they monitor
and manage.
A major initiative being pursued by NCSD CSSP to prepare for
catastrophic events against our nation's control systems is the on-
going effort to expand the US-CERT's current capability for responding
to cyber incidents and vulnerabilities to include the ability to
respond to incidents involving control systems. The NCSD CSSP provides
the US-CERT Operations Center with control system expertise and support
in responding to control system related incidents and in managing
vulnerabilities affecting our nation's critical control systems. An
important component of this US-CERT control system support is the
utilization of the knowledge, resources, and control system expertise
and cyber security expertise available among the national laboratories
and the control systems community.
NCSD is creating the infrastructure and processes to specifically
deal with both cyber attacks against control systems and also natural
disasters that affect control systems. NCSD received positive feedback
from the control system community in response to the informational
focus paper the US-CERT released to the control system community to
assist owners and operators in restarting their control systems safely
and securely in response to Hurricane Katrina. This document is
available on the US-CERT web site: http://www.us-cert.gov/reading_room/
KatrinaCSA.pdf.
On August 12, committee staff was told in a briefing with
DHS officials that there are only two full-time DHS employees working
on control system issues. How many DHS employees are currently working
on SCADA/control system issues?
Response: NCSD has authorized three government full time equivalent
(FTE) billets for the CSSP. Currently, two of those three positions are
filled and the third is expected to be filled in Q2 of FY06. In FY04,
NCSD's CSSP determined that the control systems expertise necessary for
the program to perform its mission was not readily available within the
government and sufficient authorized FTE billets were not available at
that time. In FY04, the CSSP conducted research to identify programs,
facilities, capabilities, and resources, including national
laboratories, which possess control systems and associated cyber
security expertise and resources. NCSD utilizes these identified
resources and capabilities to achieve mission goals and objectives.
The Department established the Process Control System
Forum (PCSF) to facilitate communication between government, industry,
vendors, and academia. How effective has this endeavor been? How
frequent have the meetings been?
Response: The PCSF is a relatively new endeavor and it is difficult
to assess its effectiveness at this point in time. DHS plans to conduct
an independent audit of the effectiveness of the PCSF in Q3-FY06. The
value of the PSCF is its ability to reach out to representatives from
all of these stakeholder groups in all critical infrastructure sectors
(such as chemical, water, energy and others) that utilize and rely on
SCADA/control systems. The PCSF met four times in FY05 with its next
meeting scheduled for June 6-7, 2006 in La Jolla, California
DHS has gone through four cyber security managers--Richard
Clarke, Howard Schmidt, Amit Yoran, and Robert Liscouski. How has
turnover on the DHS cyber security team impacted the effectiveness of
DHS to deal with a cyber attack? Mr. Liscouski left in January--Why
hasn?t Secretary Chertoff appointed a replacement?
Response: Addressing organizational issues is central to Secretary
Chertoff's ``Second Stage Review'' (2SR) of the Department. The 2SR
details a six-point agenda that includes improving DHS financial
management, human resource development, procurement, and information
technology, and realigning the DHS organization to maximize mission
performance. Recognizing the importance of protecting critical cyber
assets, Secretary Chertoff is increasing the authority for cyber
security by placing the coordinated activities of the NCSD and National
Communications System (NCS) under an Assistant Secretary for Cyber
Security and Telecommunications. The new Assistant Secretary will
report to the new Under Secretary of Preparedness. We expect that the
new Assistant Secretary will be named in the near future.
There are several SCADA test beds across the country. Is
there any risk of duplicating efforts with the lab beds at Sandia and
Idaho and other research? Is there any way to consolidate these
efforts?
Response: The NCSD CSSP completed an evaluation that identifies
control system security-related programs among national laboratories,
academic institutions, and agencies. This initiative evaluated the
respective value of other's work to the CSSP; and provided
recommendations on how selected program activities could be leveraged
to reduce control system vulnerabilities. The focus was on domestic
public sector programs because they could be more readily leveraged
than activities in the private and international sectors. The results
of this evaluation were utilized to identify where duplication of
efforts might exist and also served as a roadmap to identify which
groups the CSSP should work with.
The Department of Energy's Idaho National Laboratory (INL) has been
designated as the lead national laboratory in supporting the CSSP.
However, the CSSP funds initiatives with several DOE national
laboratories and the control systems community through a contract with
INL. INL has been assigned the role of coordinating and leveraging
efforts between labs utilizing specific expertise, facilities, and
capabilities at each laboratory to perform its work. In January 2005, a
Leadership Steering Group was organized, which consists of members from
INL, Sandia National Lab, Pacific Northwest National Lab, and Lawrence-
Livermore National Lab. The Group meets on a quarterly basis to discuss
the direction of the program, coordinate efforts and deliverables, and
identify expertise that is needed to solve issues and challenges. Ideas
are exchanged and security products that are developed for various
governmental customers are shared.
Moreover, utilizing more than one lab allows for additional
development and verification of efforts. If only one group is able to
address an issue, then the best achievable results are limited to what
that group develops. Competition is a motivating force that compels
people to work harder and faster to produce the greatest advances and
best solutions. Constructive competition exists among those who are
attacking SCADA/control systems, and therefore it is important to
encourage competition among those seeking to protect our systems.
This is more of a general question about fundamental
Internet protocols. There has been significant discussion in the
technology world about the security of the basic, underlying Internet
protocols. In your opinion, how secure are these protocols? Is this
something that DHS is examining?
Response: There are, and likely will continue to be, security
issues with Internet protocols. The Internet Engineering Task Force has
a Security Area, http://www.ietf.org/html.charters/wg-
dir.html#Security%20Area, which has a number of individual working
groups addressing these issues. NCSD currently does not have any
efforts or projects dedicated specifically to studying a particular
protocol, although efforts are underway within DHS to model SCADA/
control systems to better understand the disruptive effects of internet
congestion to SCADA/control systems and the effectiveness of Next
Generation Priority Services (NGPS) against these disruptions.
There is a significant challenge with the lack of security, or
verifiable security, in core internet protocols. Some application level
protocols (such as Secure Shell and Secure Socket Layer) and their
implementations have improved their security over the last few years.
However, the core security problems with underlying protocols,
transport layer and below (e.g., Transmission Control Protocol/Internet
Protocol and Address Resolution Protocol), create long term security
problems. Although some credible attempts at improving these underlying
protocols are ongoing (e.g., Internet Protocol Version 6), the question
of their overall security remains unanswered.
The National Strategy to Secure Cyberspace (NSSC) calls out the
fact that there are challenges with the existing Internet
infrastructure. As a step toward fulfilling its responsibility for
coordinating implementation of the NSSC with respect to the domain name
system (DNS) infrastructure, DHS S&T is working to deploy the DNS
Security Extensions (DNSSEC) protocol. The DNSSEC effort will enhance
the security of a fundamental element of the Internet infrastructure.
DNS is the hierarchical naming system that maps IP (Internet Protocol)
addresses to more user-friendly but structured names; the extensions to
the original protocol consist of a hierarchy of cryptographic
signatures that assure the integrity of the DNS queries by providing
origin authentication of DNS data, data integrity and authenticated
denial of existence. These measures protect against tampering in caches
and transmission and enhance the infrastructure's security, thus
contributing to increased trust in the Internet and systems, services
and markets that rely upon its secure operation. The DNSSEC protocol
has been under development for more than10 years and was approved by
the IESG in October 2004; it is awaiting final publication. The goal of
this effort is to enable all DNS traffic on the Internet to be DNSSEC
compliant. In operational terms, this goal translates into the
following ideal: Every lookup request requires and receives only
DNSSEC-validated answers. Achieving this operational goal occurs within
the framework of four principal and interrelated tracks: technical,
organizational, education and outreach, and public policy. The primary
focus of this effort is on the technical issues and process of adoption
and the organizational and outreach/ educational activities required to
achieve resolution of the technical objectives and activities. DHS S&T
has been responsible for coordination among government agencies, namely
Department of Commerce (DOC), Office of Management and Budget (OMB),
General Services Administration (GSA), Department of Defense (DOD), and
several others.
The NSSC also calls out the fact that there are challenges with the
existing Internet routing infrastructure. As a step toward fulfilling
its responsibility for coordinating implementation of the NSSC with
respect to the routing infrastructure, DHS S&T is working with
government and industry through the Secure Protocols for the Routing
Infrastructure (SPRI) program within the S&T Directorate. DHS S&T has
organized a series of workshops in the SPRI program to formulate an
approach and a roadmap for securing the Border Gateway Protocol (BGP)
in the Internet routing infrastructure. This workshop series has
brought together people from academia, research institutions,
government, and industry who have a thorough understanding of BGP
technology, of BGP use in the Internet today, and of the business of
providing internet service. Several techniques to secure BGP have been
suggested, but none has won acceptance in terms of completeness,
scalability or deployability. The workshops have been working towards a
consensus of an acceptable, deployable security technique and a
strategy for deployment. The SPRI initiative has been successful at
bringing together the major Internet Service Providers (ISPs), router
vendors, large-scale end users, government, and academia to identify a
path forward to harden the routing structure of the Internet. This has
included working with the major Internet registries, such as the
American Registry of Internet Numbers (ARIN) and Reseaux IP Europeens
(RIPE), and international participants from forward-looking countries,
such as Sweden, Netherlands, and Japan.
Relative to control systems, this issue is important because many
companies are now using standard Internet protocols to communicate
between the control room and the enterprise network. Control systems
vendors are beginning to use core Internet protocols as their bottom-
most communication mechanisms on control system local area networks.
Control system specific protocols tend to be insecure because they were
not designed with security as a dominant focus, many are proprietary
and depend on ``security through obscurity,'' and control system
protocols have generally not been exposed and stressed from a large
number of concentrated attacks from hacker groups.
In 2003, the President, as part of an initiative to
protect American infrastructure, ordered the Department of Homeland
Security to create The National Infrastructure Protection Plan. This
plan was due in December 2004. DHS released an Interim Report in
February, 2005, which was criticized by the GAO for being incomplete.
At the time the Interim Report was created, DHS pushed the due date for
the Final NIPP back to November, 2005. When will the Office of
Infrastructure Protection finalize the NIPP? What is the role of the
National Cyber Security Division (NCSD) in NIPP? What role will your
office be playing in the ``Final NIPP''?
Response: The draft NIPP Base Plan was released for final review
and comment on November 2nd, and addresses the Federal, State,
territorial, tribal, local, and private sector roles and
responsibilities for critical infrastructure protection. It will be
completed in early 2006. The 17 critical infrastructure and key
resource (CI/KR) Sector-Specific Plans (SSPs) will further detail risk
reduction strategies related to their respective critical cyber
infrastructure.
As part of NCSD's participation in the development of the National
Infrastructure Protection Plan (NIPP), NCSD is ensuring that the NIPP
Base Plan includes content to address cyber security and the cross-
sector/cross-border cyber element of CI/KR protection across all 17
sectors. NCSD also highlights cyber security concerns in an appendix to
the Base Plan that provides additional details on processes,
procedures, and mechanisms needed to achieve NIPP goals and the
supporting objectives for cyber security. The cyber security appendix
specifies cyber responsibilities for security partners, processes and
initiatives to reduce cyber risk, and milestones to measure progress on
enhancing the Nation's protection of cyber infrastructure.
After the release of the ``Final NIPP,'' NCSD will continue to work
with the relevant stakeholders to address cyber security and the cross-
sector cyber element of CI/KR protection as outlined in the draft. This
will include developing the Information Technology Sector Specific Plan
as the designated Sector Specific Agency for the IT Sector, providing
guidance to other Sector Specific Agencies to address cyber security,
and coordinating international aspects of cyber infrastructure
protection.
According to a New York Times article last week, DHS is
spending $17 million of its $1.3 billion science and technology budget
on cyber security. Committee staff was told in a briefing with DHS
officials that there are only two full-time DHS employees working on
control systems issues. Do you think the Department is devoting enough
attention and resources for cyber security?
Response: The Department is devoting significant resources and
attention to the important area of cyber security, as described in the
detailed answers to the questions above. NCSD and S&T continue to
partner effectively to produce tangible results in an area that is
constantly evolving. As described above, the NIPP provides a framework
and roadmap for progress and unites Federal, State, local, and tribal
governments and the private sector in the process for studying and
identifying solutions to mitigate cyber risk. Additionally, recognizing
the importance of protecting critical cyber assets, Secretary Chertoff
is increasing the authority for cyber security by placing the
coordinated activities of the NCSD and NCS under an Assistant Secretary
for Cyber Security and Telecommunications. The new Assistant Secretary
will report to the new Under Secretary of Preparedness. We expect that
the new Assistant Secretary will be named in the near future.
Questions for the Record from the Honorable Bennie G. Thompson for
Larry Todd
TOPIC I. THE THREAT: PROBABILITY/IMPACT OF ATTACKS ON SCADA SYSTEMS
Question: Based on all available research, how likely is an attack
on a SCADA system?
Answer: The Bureau of Reclamation has no specific statistics on
probability of attacks against SCADA systems in industry or the federal
government at large. Reclamation assumes, however, given the importance
of water and power infrastructure, that SCADA could be the target of an
attack.
Question: Based on available research, how frequently are SCADA
networks attacked?
Answer: The Bureau of Reclamation has no specific statistics on
attacks against SCADA systems in industry or the federal government at
large. Reclamation has monitoring systems in place and, to date, has
not identified any attacks against our SCADA systems throughout the
history of their operation. We believe this is due to the isolation of
our SCADA systems from the internet.
TOPICS II-IV--No questions pertain to the Bureau of Reclamation
TOPIC V. DAM SAFETY
Ouestion: Does the Bureau of Reclamation monitor only the 17 or so
dams that it has created? Or is the bureau monitoring and conducting
threat assessments to private dams as well?
Answer: Reclamation has constructed manages 471 dams, 58
hydroelectric powerplants, and other related facilities in the 17
Western states. For security purposes, Reclamation has identified 280
of these facilities as critical for completing security assessments.
Reclamation reassesses these facilities on a periodic basis. A security
risk assessment examines the threats, vulnerabilities, and consequences
of a security event at a facility. Although Reclamation has provided
some assistance to other Federal agencies, it does not monitor or
conduct threat assessments for private dams.
Question: Help me understand further the way that the control
systems at our nation's dams are connected to computers far from the
dams and what specific defenses you have put in place to protect those
communications links?
Answer: Reclamation uses leased lines and federal microwave
channels to address nearly all long-haul communications between SCADA
control centers and their outlying controlled sites. This is true of
all significant and critical SCADA communications. In some instances
UHF or radio communication hops may be employed to support less
significant SCADA functionality where data collection and low-risk
control functionality are involved. Short-haul communications employ
fiber-optic copper cabling for communication between control system
components that are widely distributed geographically. We use
federally-owned microwave-based telecommunications systems. In a few
cases, we also lease point-to-point circuits from telecommunications
companies. These SCADA communications circuits are dedicated (not
shared). Reclamation uses several protection methods including non-
Internet communications protocols and one-way communications paths. No
SCADA system communication takes place over the Internet.
Follow-up Question: Can those connections be used to open flood
gates? And if in the when the reservoirs are full, someone did that.
would there be a high probability of lives being lost? Have you had
damage estimates done at maior Federal dams? Do know how many lives
might be lost?
Answer: None of the Reclamation spillway gates under SCADA control
have a capacity greater than the safe channel capacity. Therefore, no
lives can be lost by flooding outside the safe channel capacity by the
mere operation of Reclamation SCADA systems. Instead, Reclamation
typically relies on manual, on-site operation of the gates. For the few
spillways that are operated with SCADA systems, safety measures are in
place. The safety measures in place include: remote monitoring of gate
position; control action timing relays that allow only limited raise or
lower motion based on a single control action the gate will only raise
or lower a certain percentage of its full travel based on one command);
and manual SCADA control lockouts that must be physically and
procedurally bypassed to enable SCADA control, thereby preventing SCADA
control of critical fully supervised. In addition, some gates have
limiting switches that only permit them to be moved a small amount at a
time.
From our dam safety program, we have estimates for each high and
significant hazard dam of population at risk (number of individuals
damaged including owned property) and loss of life in the event of
complete dam failure. In many cases, we also have estimates of
population at risk and loss of life in other flood situations such as
failure of gates. We would be willing to give you a secure briefing to
provide more information, at your request.
Question: We have heard the story of a hacker control of some
systems of the Roosevelt dam in Arizona, which holds 400 trillion
gallons of water. What is the worst damage that could be done there? Is
it possible to shut out on-site control? In other words, if someone
hacked the system and tried to release the water, switch off a hydro-
generator, etc., one would assume that there is an on-site, physical
override of the SCADA or Process Control System Is that true in all
cases?
Answer: It is true that, in 1994, a hacker dialed into a system
that monitored the water levels of canals in the Phoenix, Arizona,
area. This system was designed for water level monitoring only, and
investigators concluded that the hacking incident posed no threat to
safety. The story of a 12-year old hacker control of the floodgates at
Theodore Roosevelt dam in Arizona in 1998 is, fortunately, only a myth
of unknown origin.
The discharge capacity of the one powerplant unit at Roosevelt Dam
that can be controlled remotely by SCADA is small and well within the
safe discharge capacity of the downstream Salt River. Such a discharge
could also be easily handled at Horse Mesa Dam, Mormon Flat Dam,
Stewart Mountain Dam, and Granite Reef Diversion Dam, all downstream of
Roosevelt Dam. An intruder into the SCADA system cannot cause any
releases of water from the dam that will result in any downstream flood
damage or threaten the safety of any downstream populations.
SCADA control capabilities can always be disabled at the controlled
device (generator, gate, valve, etc.) via a manually operated local
control switch.
Question: Are stand-alone networks used at dams, or do you
piggyback on the local phone network, the Internet, or some other
existing outside network? Is there a Bureau of Reclamation policy on
what networks can be used for SCADA/PCS?
Answer: SCADA networks are isolated from networks other than
similar SCADA networks. Reclamation's policy addresses all networks
(including SCADA) and includes network expansions and extensions, which
must be approved by Reclamation's Chief Information Officer. Approval
adheres to guidance of the National Institute for Standards and
Technology (NIST) and is based on internal vulnerability assessments.
Question: Generally are the Cyber Security requirements of the
Bureau of Reclamation department-wide or do have different requirements
for each dam? If you have a Bureau of Reclamation Standard, is it the
same as the Army Corps of Engineers, the Tennessee Valley Authority,
and other federal agencies/entities?
Answer: Reclamation applies the same baseline cyber security
requirements to all of its systems, regardless of the type of system or
its location. In some instances, additional security requirements are
imposed because of the higher criticality or sensitivity of the
information or functions processed by a cyber system. Many SCADA
systems fall into this higher criticality or sensitivity category and
are consequently held to higher security requirements. In all cases,
however, these additional requirements are consistent with NIST and
Federal Information Processing Standards (FIPS) guidance.
Although the security foundation requirements for all federal
entities are very similar for systems of similar sensitivity and
criticality, civilian agencies, such as the Department of the Interior,
are subject to the cyber security guidance published by NIST. Agencies
under the Department of Defense, such as the Army Corps of Engineers,
are subject to a different set of policy, standards, and guidance.
Cyber security policy developed by the Department of the Interior and
the Bureau of Reclamation will probably not be identical to that
prepared by the Army Corps of Engineers, the Tennessee Valley
Authority, or other federal entities. The differences, though, are
likely to be in details related to meeting mission and organizational
needs and requirements, not in foundational cyber security requirements
or security best practices.
Question: Do all Bureau of Reclamation dams use the Risk Assessment
for Dams to assess the threat, vulnerabilities, consequences, and
ultimate risk that the faces?
Answer: Reclamation uses three methodologies depending on facility
criticality. For National Critical Infrastructures, Reclamation uses
the Defense Threat Reduction Agency assessments. For 50 of our critical
facilities, we use the RAM-D methodology. For lower priority
facilities, Reclamation uses the Matrix Security Risk Analysis (MSRA)
Question: How Bureau of Reclamation facilities have done RAM-D or
other assessments? Have those vulnerabilities been addressed so that
security is up to an acceptable level?
Answer: Following the events of 9-11, security was enhanced at all
Reclamation facilities, with full time guards and patrols being
deployed to the most critical facilities. Reclamation initiated
comprehensive security risk assessments at all 280 critical facilities,
completing the most critical facilities in 2002 and the less critical
ones this past year. The assessments identified potential threats,
vulnerabilities, and consequences. The assessments resulted in numerous
recommendations for enhancing security through both procedures and
facility fortifications. Recommendations for enhancing security
procedures were implemented upon completion of the assessments, as they
generally did not require new funding. Recommendations for facility
fortifications require additional funding, and those are being
programmed and implemented on a priority basis. Security fortifications
are complete at one National Critical Infrastructure (NCI) facility and
in progress at the other and several Major Mission Critical (MMC)
facilities. Over 73% of all recommendations resulting from the risk
assessments have already been implemented.
Question: Dams are one of the Key Asset Sectors identified in
Homeland Presidential Directive 7. Since the issuance of HSPD 7, how
much has the Bureau of Reclamation's increased? Have you had to shift
spending from other priorities to pay for security?
Answer: Reclamation's enacted and requested security budgets have
increased over the FY 2003 appropriated security budget of $28,440,000.
Reclamation continues to take its security responsibilities seriously,
and aligns security priorities with all other mission critical
programs.
Following is a brief summary of Reclamation funding for security
for Fiscal Years 2003 through 2006:
FY 2003: $28,440,000 appropriated
FY 2004: $28,583,000 appropriated
FY 2005: $43,216,000 appropriated 2006:
FY 2006: $50,000,000 ($40 million appropriated $10 million from
beneficiaries)
Responses from Dr. Sam Varnado to the Honorable Bennie G. Thompson
Questions
I. THE THREAT: PROBABILITY/IMPACT OF ATTACKS ON SCADA SYSTEMS
(To all) Based on all available research, how likely
is an attack on a SCADA system?
The probability of an attack by a dedicated adversary is not known.
The probability of nuisance acts, occurring on a daily basis, is 100%.
There is no current, reliable, classified or unclassified estimate
of the specific probability of a malevolent attack on SCADA systems.
However, we know SCADA systems are vulnerable. We also note an article
in the June 27, 2002 Washington Post that these systems have been
targeted by al-Qa'ida terrorists who have a great deal of capability
and patience. There are signs that hacker coalitions and nation states
are collecting information on SCADA systems. The sophisticated threats
have significant financial resources and can attack at will. Because of
the commonality of computing platforms in a networked system, an attack
that is successful against one will almost surely succeed against them
all, and at only slight additional cost to the attacker.
SCADA systems are now moving from the old stand-alone legacy
systems to systems that use the internet or local enterprise networks
as the backbone. This means that all the current computer attack
modes--worms, viruses, denial of service-can now deny or disable
control systems. It is no longer a requirement for a successful
attacker to be a control systems expert to bring down a SCADA system.
These types of attacks occur daily.
(To any of the labs) What cyber security failures and
incidents have you seen with SCADA networks?
Sandia National Laboratories has performed numerous critical
infrastructure assessments that identified common vulnerabilities in
SCADA systems. The results are published in a paper entitled ``Common
Vulnerabilities in Critical Infrastructure Control Systems'' that can
be found at http://www.sandia.gov/scada/documents/031177C.pdf. This
paper describes the types of vulnerabilities we have identified.
In addition to our assessments, there have been the following
documented incidents:
It has been reported that in June 1982, exploitation of SCADA
software created a damaging attack on the Trans-Siberian
pipeline. The software that was used to run the pumps,
turbines, and valves of the pipeline was programmed to
malfunction after a specific time interval. The malfunction
caused the control system to reset the pump speeds and value
settings to produce pressures beyond the failure ratings of the
pipeline joints and welds. The result was the largest non-
nuclear explosion (3 kilotons) ever seen from space.
In January 2003, the ``Slammer'' worm disabled a monitoring system
at the Ohio Davis-Besse nuclear power plant. The worm entered through
an improperly secured network connection to a contractor's facility.
The worm crashed the computerized panel used to monitor the plants most
crucial safety indicators. This incident did not pose a safety threat
at the time because the reactor was offline for repairs and the
redundant analog monitoring systems were still in operation. However,
this event illustrates the impact that a computer worm can have on a
SCADA System. Reference: ``Slammer worm crashed Ohio nuke plant
network'', Kevin Poulsen, Security Focus (19 august 2003): http://
www.securityfocus.com/news/6767
In May 2001, attackers were apparently able to gain access to one
of the computer networks at the California Independent System Operator
(Cal-ISO) corporation. This hacking incident was apparently
unsuccessful at penetrating any process control system network, yet it
uncomfortably extended over a period of more than two weeks. Reference:
``California hack points to possible IT surveillance threat,'' Dan
Verton, Computerworld (12 June 2001): http://www.computerworld.com/
industrytopics/energy/story/0,10801,61313,00.html
One verified attack occurred in April 2000 at Maroochy Shire,
Queensland. Disruption of the SCADA systems that controlled the plant
resulted in release of copious quantities of sewage into parks, rivers,
and a hotel, severely fouling the environment. Reference: ``Hacker
jailed for revenge sewage attacks,'' Tony Smith, The Register (UK) (31
October 2001): http://www.theregister.co.uk/content/4/22579.html
At about 3:28 PM Pacific Daylight Time on June 10, 1999, a 16-inch-
diameter steel pipeline owned by Olympic Pipe Line Company ruptured and
released about 237,000 gallons of gasoline into a creek that flowed
through Whatcom Falls Park in Bellingham, Washington. About 1.5 hours
after the rupture, the gasoline ignited and burned approximately 1.5
miles along the creek. Two 10 year-old boys and an 18-year-old young
man died as a result of the accident. Eight additional injuries were
documented. A single-family residence and the city of Bellingham's
water treatment plant were severely damaged. As of January 2002,
Olympic estimated that total property damages were at least $45
million. The National Transportation Safety Board listed five reasons
for the rupture. The fifth was Olympic Pipe Line Company's practice of
performing database development work on the SCADA system while the
system was also being used to operate the pipeline, which led to the
system's becoming nonresponsive at a critical time during pipeline
operations. Reference: http://www.ntsb.gov/publictn/2002/PAR0202.htm
(To all) Based on all available research, how frequently
are SCADA networks attacked?
Again, the answer depends in part on how one defines ``attack''. If
attack includes active scanning, attempts to take advantage of
unpatched vulnerabilities, worms, viruses, and spyware, then any
control system network connected directly or through a business network
to the Internet is under constant attack. It is reasonable to assume
that network-connected SCADA systems across the country are probed
daily.
There have not been many documented malevolent attacks of SCADA or
control systems. Attacks do happen, and there are more attacks then we
know about because some infrastructure owners are reluctant to report
SCADA attacks. They worry about loss of public confidence and
competitive issues. We have seen a few targeted attacks in our 10 years
of experience.
(To any of the labs) Is it possible to devise an attack to
disable or disrupt a SCADA network for an extended period of time? If
so, what is being done to mitigate such attacks?
Yes, it is possible to disable or disrupt a SCADA network for an
extended period of time. The exact method of attack depends on the
individual circumstances of the SCADA network. The Maroochy Shire
wastewater SCADA system attack in Australia is often cited because the
details are unclassified. Whether one considers the consequences
significant or not, the fact remains that disgruntled computer expert
Vitek Boden caused a chronic disruption of a SCADA network for three
months. His attack could have been more sophisticated and, possibly,
might have caused greater consequences. More significantly, the SCADA
components he attacked are commonly used in domestic water treatment
systems. Sandia's internal research and development has discovered
forms of attack that could result in even greater consequences. The
details of these attacks are classified and would need to be shared in
a different venue.
The responsibility for mitigation is distributed among the SCADA
network owner/operators, the SCADA network integrators, the SCADA
equipment vendors, industry groups, and regulators. Even when one of
the players takes responsibility for security, they can only mitigate
the portion they control. Operators can put in place security policies,
plans, and implementation, but they are at the mercy of vendors who may
not provide features necessary for security. For this reason, the
degree of mitigation of SCADA networks is highly variable.
Mitigation effects may not be implemented for several reasons.
First, a business case for industry to invest in SCADA security has not
been clearly made. As a result, funding for security personnel and
equipment are often inadequate.
A second problem is natural attrition through aging of key
personnel in utility operations. Taken together, it is probable that
quick automation repairs will no longer be possible for many utilities
in the very near future, primarily because of a shortage of trained
personnel and old equipment. Backup manual operation is further
exacerbated by the paucity of skilled and experienced personnel. There
are also limitations on the number of field operators, to deploy to
remote locations in manual situations when data are unavailable to the
SCADA system. Therefore, if the loss of some automation functionality
will likely cause severe problems for utility operations (including
system management functions, system/plant automated control, or any of
the supporting data categories), a redundant system and/or network is
required.
Third, classification, anti-trust, and proprietary issues get in
the way of the open sharing of threat and vulnerability information
among industry stakeholders.
Sandia has been teaching courses on SCADA security assessment and
best practices for mitigation to industry and government for several
years. In that time, our message has been heard by some entities, who
are now asking for more information. We have performed vulnerability
assessments that continue to confirm the presence of common
vulnerabilities.
(To KP Ananth or Sam Varnado) Electric power is important
for nearly all the things that Americans do--from businesses to schools
to government to many forms of recreation. Has your research shown that
the SCADA systems that control our power generation and distribution
are fully protected from attacks launched from the Internet? If not,
what kind of damage do your researchers believe smart, well researched
attacks could cause?
SCADA networks that control electric power generation,
transmission, and distribution are not fully protected from attacks
launched from the internet. Well researched attacks can cause burn-out
of expensive, hard-to-replace equipment such as transformers. The
duration of such outages could extend to several months. Other computer
attacks, such as worms or viruses, could create outages lasting for
days.
Further information about the consequences of a smart, well-
researched attack is available at a classified level and could be
provided in another venue.
(To Sam Varnado, KP Ananth, Bill Rush) We've heard a lot
about the impact of a terrorist attack on a control system. But as we
saw during Katrina, natural disasters can cause devastating impacts to
our control system infrastructure too. What kind of impact would a
natural disaster have on control systems in California (earthquakes),
Oregon (tidal waves/tsunamis), the Gulf Coast (hurricanes), and
elsewhere?
Terrorist attacks differ from natural disasters in that the
terrorists take a functional attack perspective. In other words, they
look to destroy or alter the functionality of a SCADA system. In
contrast, a natural disaster is random and geographically dependant.
Anything within the physical range of the disaster is affected.
Anything outside is less likely to be affected. Many companies have
created redundant control centers to better prepare for such disasters.
The critical assets are identified and duplicated, and risk-mitigation
plans are usually in effect.
In some respects, certain natural disasters are easier to handle
than focused cyber attacks. A crew made up of control specialists and
physical facilities members can very quickly determine what physical
assets have been damaged. These assets can be reordered and replaced
like any other field equipment. Typically, control systems are composed
of off-the-shelf parts and reordering is not usually a problem.
Lack of warning is one aspect that makes response to some disasters
more difficult. Hurricanes are different than earthquakes, tsunamis,
and terrorist events. Damage can be minimized if there is enough
warning to allow shut down. When the event happens with little or no
time to prepare, the chance for damage increases. Listed below are the
areas of concern, the disaster being considered, and the potential
impact.
Gulf Coast:
Natural Disaster: Hurricane
Infrastructure: Oil, Gas, chemical, electrical
Impact: Because of pre-warning, these infrastructures are reasonably
well equipped to deal with the disaster. Control system equipment can
be damaged or destroyed, resulting in outages of service. However, if
the infrastructure elements are shut down prior to the storm, damage
can be minimized.
California:
Natural Disaster: Earthquake
Infrastructure: Oil, Electricity, Telecommunication, Natural Gas
Impact: Without warning, many of the infrastructure control systems
could be severely damaged through physical destruction of computer
facilities. Impacts could be severe and widespread. However, backup
systems located in unaffected areas will help minimize the impact and
help in system recovery.
Oregon:
Natural Disaster: Earthquake, Tidal Wave
Infrastructure: Oil, Electricity, Natural Gas
Impact: Tidal waves are of less concern than earthquakes. Most
infrastructure assets are well protected from tidal waves by
landmasses, but they lie in a critical area for earthquakes. Loss of
electricity because of extensive physical damage could lead to failures
in other infrastructures because they need electricity in order to
safely shut down. In addition, the economy in the pacific Northwest
could be severely impacted if electrical failures caused a disruption
of port activities.
The Department of Homeland Security's (DHS's) Infrastructure
Simulation and Analysis Center (NISAC) at Sandia has created a number
of relevant reports on the economic consequences of natural disasters
as follows:
Numerous Katrina reports on damage from Katrina both
before and after land fall
A report entitled ``Infrastructure Assets in
Seismically Active Zones in the Pacific Northwest''; this
report addresses assets located in Washington, Oregon, and
Idaho
Analysis of economic impacts of port disruptions in
the Pacific Northwest.
Natural disasters affect all critical infrastructures. The
interdependent nature of the infrastructure amplifies the consequences
of disruption in any one sector. Fortunately, preparing for the
abnormal natural disaster event also helps prepare for the malevolent
attack. Many of the practices that Sandia teaches in our course on
sustainable security are equally applicable to sustaining operations
during natural disasters and recovering after those disasters.
II. THE PUBLIC/PRIVATE RELATIONSHIP IN DEVELOPING A SCADA SOLUTION
(To any of the labs) I understand the National Labs
are conducting extensive research into SCADA and control
systems. What resources are you currently lacking? How are you
coordinating these efforts with the private sector? What can
the federal government do to provide you with more resources?
Our biggest need is predictable, sustainable, multi-year funding
tied to a well-defined research and development plan. We have
outstanding well-trained staff who are experts in cyber security.
However, cyber research has not been emphasized by DHS. DHS should
ensure that the best technical capability in the country is applied to
this problem. The national labs--particularly Sandia and Idaho national
laboratories--have the necessary talent, but DHS needs more funding to
apply to the problem.
In addition, existing DHS programs, emphasize the conventional
hacker threat. There is a need to address the more sophisticated
threats such as those coming from terrorists and nation states. Sandia
has outstanding capabilities in these areas, but they are not being
applied to the SCADA problem.
How are you coordinating these efforts with the
private sector?
We are currently working with DOE and private industry to develop a
roadmap for securing the nation's energy infrastructure from the cyber
threat. In addition, we currently engage in a variety of outreach and
awareness activities, including teaching vulnerability assessment and
SCADA security courses to industry, making technical presentations, and
providing the products of our research on a website, http://
www.sandia.gov/scada/. We participate in programs such as the Institute
for Information Infrastructure Protection (I3P), Linking the Oil and
Gas Industry to Improve Cyber Security (LOGI2C), Process Control
Systems Forum (PCSF), and the National SCADA Test Bed (NSTB); all are
aimed at fostering cooperation and coordination with industry. We also
frequently host visits from industry to Sandia.
Additionally, we provide training on risk assessment methodology
and vulnerability mitigation to a wide range of industrial customers.
What can the Federal Government do to provide you with
more resources?
Funding should be increased for improvements in cyber security
technology so that DHS can provide tools for
high speed intrusion detection systems
software assurance
attack attribution and trace-back
security modeling of existing and proposed SCADA
systems
network visualization for mapping cyber disruptions
triage of threat scenarios across many vectors
assuring the reliable performance of commercial off-
the-shelf (COTS) products. We need funding of $15M/yr to apply
to this problem
models and simulations to understand the large-scale,
transient consequences of attacks on the power grid.
Funding for a new program to address the sophisticated threat
should also be provided. We anticipate that more sophisticated and
strategically integrated cyber attacks--such as those that might be
marshaled by a well-funded and highly capable terrorist or nation-state
actor--will occur against control systems. An effort is needed to
develop the analytic resources and technologies required to detect and
predict these threats based on control system vulnerabilities, to
strengthen our preventive measures, to increase our ability to respond
expediently, and to model these more sophisticated threats and analyze
the operational impacts they have on control systems. In general, this
is a better role for national laboratories than for universities and
private industry vendors. Sandia could lead this program. This effort
should include a strong emphasis on the problems of building trusted
systems from untrusted COTS components.
Further, we need funds to work more closely with industry to
provide in-depth vulnerability assessments of existing systems, to help
industry utilize existing risk assessment models, and to formulate a
business case for investment in cyber security.
Finally, DHS needs to identify the commonalties in SCADA systems
across all infrastructure elements and then define and coordinate
efforts for improving SCADA system security across these
infrastructures. Industry infrastructures owners should be provided a
single point of contact for their interactions with DHS.
(To any of the labs) It has been widely reported that
both industry and the federal government find it difficult to
estimate the economic impact of a cyber security attack. Has
the lack of actual quantifiable damages made the private sector
leery of investing in cyber security?
At the I3P SCADA Security Conference in June 2005, held in Houston,
panelists from industry made exactly this point. They said:
``The lack of quantifiable damages is one of the missing
components that would feed into the private sector's cost-
benefit and return-on-investment analysis. The economic case
for investing in cyber security has to be stronger than the
economic case for investing in anything else before the private
sector will be compelled to make cyber security investments.''
This observation illustrates the difficulty that industry is having
in making a business case for investment in cyber security. There are
two steps that will help overcome the noted deficiency. First, DHS
should fund the national laboratories to work with industry in
utilizing the lab's risk assessment methodology to help industry make
the business case. Second, DHS should apply the skills of NISAC, run by
Sandia and Los Alamos labs, to the problem of determining the economic
consequences of infrastructure outages caused by cyber attacks.
(To Sam Varnado, KP Ananth, Bill Rush) Can you tell us
specifically how your research on SCADA has, to date, impacted
the way SCADA systems in the field are secured, and what
percentage of those systems have been impacted? If that's not a
big number, what is stopping us from putting the results of
your research into practice in the field?
We have directly affected relatively few systems, on the order of
tens. Unfortunately, our program is small and the number of control
systems is huge. We have indirectly affected--either by developing
self-assessment methodologies or through outreach--on the order of
hundreds of control systems. We have diffused our standards work to
thousands of control systems. In spite of such efforts, we have only
affected a small fraction of the control systems on which the nation
depends for its current infrastructure security.
The biggest obstacle to technology transfer is the business case
issue. Even when industry believes there is a business case for
security measures, they believe that they need only increase security
enough to protect against the low-level threat--background noise,
individual hackers, and possibly hacktivists. It is industry's
contention that government should protect against the larger threats--
organized crime, terrorists, and nation-state threats--either through
law-enforcement or national defense. We need to expand our public/
private partnerships to define best industry practices as a function of
risk and cost, then develop and disseminate the appropriate technology.
(To Sam Varnado, KP Ananth, Bill Rush) What has the
money we have already spent on SCADA research done to improve
SCADA security in the field?
A specific instance of improved SCADA security is the work
conducted to develop RAM-W, a self-assessment methodology for water
utilities. Hundreds of water utilities used that methodology to help
secure their SCADA systems. One particular utility, Washington
Aqueduct, operated by the US Army Corp of Engineers, has benefited
directly from the assessment and the secure design requirements that
Sandia provided for their new SCADA system as a follow-on project.
We have been active in international standards organizations by
helping to provide a security perspective to their guidelines, by
developing training classes, and by developing self-assessment
methodologies. We have also developed technology to secure
communication links and improve cryptographic research.
We have published and distributed to industry a report entitled
``Common Vulnerabilities in Critical Infrastructure Control Systems.''
We have also provided training courses to industry on vulnerability
assessments of SCADA systems as well as risk assessment methodologies
to help industry solve its own problems.
Further, we have identified specific vulnerabilities in SCADA
systems from several vendors. We have also explained to those vendors
how the vulnerabilities can be mitigated.
Over the last ten years, Sandia has invested in SCADA security
research, through its own internal research and development funds, on
the order of $4 million. Currently we are funded through external
sources--DHS, DOE, industry, and university collaborations--at
approximately $3 million this fiscal year. This level of funding is not
adequate to address the very hard problems that SCADA security
presents.
(To Sam Varnado, KP Ananth, Bill Rush) Is there any
risk of duplicating efforts with the lab beds at Sandia and
Idaho and other research around the country?
There is no duplication. The efforts are complementary, with each
lab applying its unique capabilities to different parts of the
problems.
The test bed at Idaho National Laboratory is designed to
demonstrate the effects of cyber attacks on large scale physical
structures. It is a unique facility.
The test bed at Sandia is in reality a SCADA security laboratory
that conducts leading-edge research on cyber security methods such as
vulnerability assessments, cryptography, security of wireless networks,
and threat analysis. It provides the capability to test the robustness
of SCADA systems from various vendors in a laboratory environment at
low cost. It is also set up to evaluate the more sophisticated
adversaries.
Further, DHS manages the work at both labs and provides a program
manager to make sure tasks are assigned in a way that avoids
duplication. It is important that DHS understands and acknowledges the
uniqueness of each lab and works to make sure that the participants at
one lab do not duplicate existing capabilities at the other lab.
III. THE FEDERAL GOVERNMENT'S ROLE IN CYBER SECURITY
(To Andy Purdy and ALL) There are several SCADA test
beds across the country. Is there any risk of duplicating
efforts with the lab beds at Sandia and Idaho and other
research? Is there any way to consolidate these efforts?
See our answer on duplication under the preceding question.
Consolidating these facilities does not make sense because they
have separate roles. One is a large, full-scale test and demonstration
facility; the other is a state-of-the-art research facility needed for
developing countermeasures for the increasingly sophisticated threat
environment.
IV. THE FEDERAL ROLE IN THE FUTURE
(To Sam Varnado and K.P Ananth) Based on your
knowledge of the SCADA research field, what are the most
promising technological breakthroughs you see that can protect
our SCADA systems in the short term? I realize there are no
silver bullets, but please list the solutions that will
actually work to protect our SCADA systems.
First, industry infrastructure owners need to define security
policies and best practices for their own systems. Security is not just
a technology problem. It is one of sustainable security--hardware,
software, people, and procedures. Employees need to be trained in
detecting attacks. Widespread adoption of best security practices has
high payoff and low costs. If all control systems implement best
security practices, the bar will be raised against all adversaries.
Second, the latest security advances such as intrusion detection
systems, firewalls, encryption, and other technologies should be
applied. For example, the application of new Layer 3 firewalls in
switches is emerging and shows promise for improving the security of
control systems.
Third, vulnerability assessments need to be performed on all major
SCADA systems. Then the identified vulnerabilities need to be
mitigated.
Finally, a strong, sustainable R&D program needs to be implemented
to continue to develop technology for countering new, more
sophisticated threats by hackers and cyber terrorists who change their
attack methods on a very frequent basis.
(To any of the labs) How do we make rapid progress in
improving security in the field?
We must help infrastructure owners develop security policies and
train their people.
We must provide incentives and liability relief to developers and
adopters of security technology. The Safety Act is a good step in this
direction.
We must support more research into robust, distributed,
introspective systems; more research into secure operating systems;
and--to achieve a high level of security--implement a dedicated
internet protocol (IP) and a redesigned IP stack for SCADA use only.
We must enable greater access to, and partnerships among, vendors,
labs, and asset owner/operators in order to better understand industry
facilities, processes, and more technology from the labs to the field.
We must provide better and clearer communication among
organizations working on cyber security to help us develop consensus on
the best security solutions. We must also promote opportunities to
provide awareness and training to vendors and asset owner/operators.
(Any of the labs) Has the federal government advocated for
standards establishing a minimum floor for securing control systems?
What would a minimum floor look like? Have industry leaders begun the
process of developing those standards already? Has the government
established any ``best practices'' that can be modeled by industry?
What other standards activities are being developed besides AGA 12?
To our knowledge, three government initiatives exist today to
address securing control systems by providing guidelines and/or cyber
security requirements to industry: (a) the Technical Support Working
Group (TSWG) ``Securing Your Industrial Control System'' guide book;
(b) the NIST release of the ``Guide to Supervisory Control and Data
Acquisition and Industrial Control System Security''; and (c) the DHS
US-CERT Control Systems Security Center (CSSC) Program cyber security
protection framework, which includes a set of cyber-security
requirements planned to be released in 2006. Whether these individual
government released documents constitute a ``minimum level of
standards/guidelines'' is not clear.
From our experience, a minimum set of security control system
standards would not come from a single standards body but would most
likely comprise the work of various standards bodies. There is no
single standards body to provide a comprehensive list of control
systems cyber security standards.
Industry-led standards bodies have begun developing standards to
address the issue of securing control systems. However, dozens of
groups/organizations currently exist that are working on control
systems security standards. Coordination of these efforts is both
essential and, at the same time, difficult. Inconsistent and
conflicting standards generated from these various groups confuse
industry and asset owners/providers. A more concerted effort on the
part of the government is needed to assist industry and asset owners in
(1) maneuvering through the abundance of control systems cyber security
standards and (2) encouraging them to develop consistent control
systems cyber security standards across all critical infrastructure
sectors. A single point of contact within DHS for cross-sector
involvement in control systems cyber security standards is needed. This
point of contact would facilitate and assist in directing industry
partners to relevant security guidelines, practices, and standards, and
it would encourage consistent application of cyber security standards.
Other standards bodies include API 1164, CIDX, FIPS Pub 200, ISA
SP99, NERC, and NIST SP800-53--as well as others too numerous to list.
The international standards bodies (e.g., IEC) are an important group
because the majority of SCADA vendors are international and follow
those guidelines.
(To any of the labs) Some have mentioned the value of a
``vendor'' incentives system that would provide tax and other financial
incentives to manufacturers who are producing control systems that are
already in ``best practice'' compliance. How feasible is this, and have
there been evaluations of the cost to the federal government?
Best practice compliance can be conducted at a component or sub-
system level if clear metrics are established to define the practice.
But even here care must be taken not to impose a standard on something
that a later technology might supersede. Cyber security technology is a
rapidly changing field.
Great care would need to be taken to insure that the ``best
practice'' standards would not be negotiated down to the point that
companies just need to fill out the right forms and jump through the
right legal hoops--doing little to actually improve security. A third
party, Underwriter's Laboratory approach may be necessary to properly
evaluate vendor's products and validate claims. Some analysis should
also be performed to determine the appropriate incentives for
compliance (industry, company, product, etc.).