Government and private sector functions depend substantially upon information and communication technology.
Cyber threats are posed to information and communications technology (ICT) systems operated by the federal government and by its contractors. Federal interests are in jeopardy if sensitive government data, residing in or transiting through such systems, are destroyed, compromised or stolen. Consequences include impairment of government and private sector functions and loss, corruption or improper use of sensitive and proprietary data.
A Vulnerable Supply Chain
The ICT supply chain is a:
complex, globally distributed, and interconnected ecosystem that is long, has geographically diverse routes, and consists of multiple tiers of outsourcing. This ecosystem is composed of public and private sector entities (e.g., acquirers, system integrators, suppliers, and external service providers) and technology, law, policy, procedures, and practices that interact to design, manufacture, distribute, deploy, and use ICT products and services.
The security of federal information often depends upon measures taken by its contractors (and their suppliers). That the federal government in 2011 adopted a “cloud first” policy further divests federal agencies of direct authority over systems that host, transmit or employ federal information.
https://cio.gov/wp-content/uploads/2012/09/fedrampmemo.pdf.
The ICT supply chain has many points of vulnerability. While the threats differ and the attack vectors are diverse, vulnerability is present at levels that extend to individual electrical, electronic or electromechanical parts as well as electronic assemblies, systems and networks. Areas that may be vulnerable to hostile cyber acts include hardware, where electronic parts exercise control functions, as well as firmware and software.
The global nature of the information technology supply chain contributes to the proliferation of these risks. Because of omnipresent interconnection, and increasing use of information services that depend upon externally managed services, cloud infrastructure and Web-enabled delivery, threats to information systems may be directed at the “weakest links” of connected enterprises.
Federal agencies employ a variety of controls to protect sensitive information when it is within the domain or authorization boundary of “federal information systems.” But vast amounts of federal information are constantly in the hands of the external federal supply chain. As to this wealth of information, in commercial systems, the presence of security controls is problematic at best.
With limited exceptions,
Corresponding action has not yet been taken on the civil side of federal contracting. Yet, the commercial supply chain that supports federal civil functions is exposed to substantially the same or similar risks. Federal agencies apply a variety of cybersecurity controls to contractors who operate ICT as “federal information systems.”
http://tinyurl.com/n4rnqkj 223 Privacy Law Watch, 11/22/10, 9 PVLR 1592, 11/22/10. The executive order states as its purpose to “establish a uniform program for managing information that requires safeguarding or dissemination controls.” The National Archives and Records Administration (NARA) is the executive agent assigned to implement
Protecting CUI—Crucial Questions
The National Institute of Standards and Technology (NIST) is working now to complete Special Publication (SP) 800-171, a control regime to protect CUI on nonfederal information systems.
http://csrc.nist.gov/publications/drafts/800-171/sp800_171_second_draft.pdf 14 PVLR 657, 4/13/15, 66 Privacy Law Watch, 4/7/15.
The first is definitional. For years, the federal government has struggled to reconcile conflicting definitions of CUI.
NARA seeks to control the designation of CUI. Its proposed rule asserts that agencies “may not control any unclassified information outside of the CUI program.” NARA would allow agencies to designate CUI only if approved by NARA as the CUI Executive Agent.
A related question is how agencies and their suppliers will be informed that information has been categorized as CUI. The proposed NARA rule asserts that agencies “must mark CUI” according to “marking guidance” that NARA will issue.
A further and crucial question is what level of safeguarding will be applied for CUI or the various categories and subcategories articulated in the CUI Registry. NARA’s proposed rule sets a “Basic” safeguarding standard, which it describes as “the default set of standards agencies must apply to all CUI” unless the CUI Registry specifies otherwise.
Industry may welcome the objective of standardization of safeguards for federal agencies to insist upon when it comes to handling their CUI. Whether this will come to be as NARA proposes is dubious. Agencies are in the best position to identify information in their domain that merits protection against cyber threats. They may be in the best position to know where vulnerabilities exist in external (contractor) information systems that expose their CUI to cyberattack. Agencies also are best able to assess the impact to their operations should the confidentiality (or integrity) of their CUI suffer compromise. All these considerations suggest that agencies will insist upon greater authority over the level of safeguards to be applied to companies that possess the agencies’ CUI. While the NARA rule provides for a special and potentially more demanding level of controls for CUI Specified, today the CUI Registry does not specify what those controls are or might be, and there is reason to doubt that agencies will acquiesce to having NARA decide in all cases on whether more controls or needed or what those controls will be. At the same time, agencies also should be aware that elevating controls beyond the CUI Basic norm can produce potentially dysfunctional consequences. Some capable and trustworthy commercial suppliers will refuse to adopt special and more demanding controls. Some companies may exit the federal marketplace altogether if they cannot reconcile special federal cyber control obligations with their general enterprise systems, or if the costs cannot be recovered. Isolated, contract-specific or CUI category-specific controls that increase contractor costs will mean higher costs to agencies that demand them. For all these reasons, and more, each federal agency should weigh carefully whether to impose cyber control requirements beyond what NARA envisions and what NIST will recommend in SP 800-171. Still, at this juncture, neither agency nor industry can foresee whether the federal CUI cybersecurity initiative will produce a generally applicable set of commercially acceptable cybersecurity safeguards, with limited exceptions, or a patchwork of specialized and different demands that industry will find very burdensome.
Boundary-Setting Problems
Tough questions are present as the federal government seeks to define and designate CUI and to establish a mechanism to require its safeguarding that is at once workable in commercial contractor environments and sufficient to respect bona fide agency concerns. A further analytic problem is present in the pervasive difficulty, in an interconnected world, of setting control boundaries for “federal information systems” as distinct from “nonfederal” or “external” information systems.
SP 800-171 evidences effort by NIST to reconcile its controls with other regimes and methods already employed in the private sector. The importance of the proposition is difficult to overstate. Federal agencies ultimately will pay the costs of mandatory cybersecurity measures imposed upon the federal supply chain. Those costs may be higher prices for supplies or services or lost access to sources that choose not to accommodate the federal demands. Means must be found to achieve the objectives of NARA and NIST to better safeguard CUI without denying or superseding the validity of other strategies and techniques as may be sufficient, but different.
Cyber Threats to Federal Information—and Citizen Privacy
Cyber threats are very much in the public mind. Most of the publicized attacks have been against the private sector. The hack of Sony Pictures Entertainment Inc. brought down that company’s information systems and disrupted day-to-day operations, while the release of supposedly “private” information caused great embarrassment. The attack on Anthem Inc. apparently compromised the health-care information of millions of insured persons. A recently reported cyber theft suggests that hundreds of millions of dollars were stolen from as many as 100 banks (or more) in the U.S., European Union and Russia. Those attacks warn that similar vulnerabilities are present in commercial ICT systems that host or act on federal information—with comparable (or worse) adverse consequences. Civilian federal agencies are responsible for CUI equal to or more sensitive than that taken from Anthem. They preside over funds even larger and financial functions even more important than those exposed by the bank cyber theft.
That CUI includes information that implicates important confidentiality interests of both individuals and our government is well stated in NIST SP 800-171:
Many federal contractors, for example, routinely process, store, and transmit sensitive federal information in their information systems to support the delivery of essential products and services to federal agencies (e.g., providing credit card and other financial services; providing Web and electronic mail services; conducting background investigations for security clearances; processing health care data; providing cloud services; and developing communications, satellite, and weapons systems). Additionally, federal information is frequently provided to or shared with entities such as State and local governments, colleges and universities, and independent research organizations.
There is official recognition of the serious and growing threat to government systems. The Government Accountability Office (GAO) has just released a report to Congress with this very disturbing summary:
[C]yber threats and incidents to systems supporting the federal government and national critical infrastructures are increasing. These threats come from a variety of sources and vary in terms of the types and capabilities of the actors, their willingness to act, and their motives. For example, advanced persistent threats—where adversaries possess sophisticated levels of expertise and significant resources to pursue their objectives—pose increasing risks. Further underscoring this risk are the increases in incidents that could threaten national security, public health, and safety, or lead to inappropriate access to and disclosure, modification, or destruction of sensitive information. Such incidents may be unintentional, such as a service disruption due to an equipment failure or a natural event, or intentional, where for example, a hacker attacks a computer network or system. Over the past 8 years, the number of information security incidents reported by federal agencies to the U.S. Computer Emergency Readiness Team (US-CERT) has increased from 5,503 in fiscal year 2006 to 67,168 in fiscal year 2014, an increase of 1,121 percent.
This report confirms that the cyber threat extends to federal information systems
Using Acquisition Planning and Contract Administration to Improve Contractor Cybersecurity
Several regimes are in place for cybersecurity and information assurance for federal information systems. These include the Federal Information Systems Management Act (FISMA),
http://www.gsa.gov/portal/content/150159.
http://cloud.cio.gov/fedramp.
http://www.whitehouse.gov/omb/circulars_a130.
http://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-53r4.pdf 85 Privacy Law Watch, 5/2/13, 12 PVLR 774, 5/6/13.
The practices, controls and standards that ostensibly apply to federal information systems, however, do not now regularly extend to nonfederal information systems. The boundaries between “federal” and “nonfederal” information systems are difficult to distinguish.
http://csrc.nist.gov/publications/nistpubs/800-37-rev1/sp800-37-rev1-final.pdf 44 Privacy Law Watch, 3/9/10, 9 PVLR 362, 3/8/10. Cloud information services are delivered through the use of “external information systems.”
Through issuance of
http://www.gpo.gov/fdsys/pkg/FR-2013-02-19/pdf/2013-03915.pdf 31 Privacy Law Watch, 2/14/13, 12 PVLR 257, 2/18/13.
Federal Market Power
That the federal government is expected to spend $90 billion on information technology (IT) in FY 2016 suggests it has market power sufficient to steer its supply chain to improve cybersecurity measures. Similarly, the very large companies that often control or operate critical infrastructure also should have sufficient influence over their supply chain to obtain improved cyber and supply chain protection.
DoD, which controls the most discretionary spending of any federal agency, already is using its contracting power—“acquisition planning” and “contract administration” measures—to improve supply chain risk management of the defense industrial base. Defense Federal Acquisition Regulation Supplement (DFARS) regulations on unclassified controlled technical information (UCTI) use acquisition methods (contract clauses and flow-down requirements) to impact all companies in the DoD supply chain.
Federal civilian agencies are working to follow suit. Shortly after issuance of
http://tinyurl.com/kfuau32 24 Privacy Law Watch, 2/5/14, 13 PVLR 212, 2/3/14.
Prudent companies should now anticipate that the federal government will use acquisition and contract tools to require commercial suppliers to improve their cybersecurity measures. Some may question whether such federal “intervention” is necessary. Market forces (and enterprise self-protection) already motivate many in the federal supply chain to improve cyber supply chain measures. No doubt, some supply chain participants will seek competitive advantage by being early adopters of more rigorous controls. However, several considerations suggest that the federal government will not trust market forces or let industry proceed at its own pace. These include the direct risk to federal interests should the confidentiality of sensitive federal information be lost by reason of cyber breaches. Recent events in the private sector vividly demonstrate the costly and lasting injury that is the consequence of a successful cyberattack, even upon supposedly well-protected systems engaged in sensitive areas of commerce. The combined efforts of NARA and NIST have as their common purpose protecting sensitive federal information (CUI) against attacks of the same or similar character.
NIST SP 800-171
NIST SP 800-171 was first released in draft, for comments, Nov. 18, 2013.
Importantly, FIPS 199 makes the determination of security categorization also a function of the “information type,” distinguishing among such types as “public information” (nonsensitive) or “investigative information” (very sensitive), and then examining impact as to each of the three security objectives (confidentiality, integrity and availability).
Three Assumptions
As presently proposed, SP 800-171 makes three express “assumptions” that explain, albeit briefly, certain “departures” from the principles that apply to CUI on federal information systems.
The support for these assumptions is questionable. Even if the requirements for the protection of CUI are the same, in fact the safeguards that NIST proposes to protect CUI, as set forth in SP 800-171, differ between federal and nonfederal information systems. Federal information systems are subject to the controls and enhancements specified in SP 800-53. For CUI in private hands, NIST proposes requirements that echo the purposes of SP 800-53 but are avowedly more accommodating of different methods. The stated assumption of SP 800-171 that the “impact” value of CUI is no lower than “moderate” likely reflects NIST’s practical recognition that contractors will object to obligations that would require different control methods to distinguish between “low” and “moderate” impact. At the same time, the assumption of common impact can be criticized as commoditizing the many variations of CUI (recognized in the NARA Registry) and homogenizing actual differentials in true impact.
The departures from the FIPS norms and separation from SP 800-53 control baselines are “concessions” that NIST has made to improve the prospects that commercial organizations (and other CUI holders, such as state and local governments and educational institutions) will accept an extension of federal controls outside their present domain of federal information systems. This is commendable restraint, but not without some “losses in translation.” For example, for federal information systems FIPS 199 and FIPS 200 work in combination. FIPS 199 distinguishes among three categories of security objectives—confidentiality, integrity and availability. It also calls for assessment of the impact of a “breach” event, again using three categories, in this case “low,” “moderate” and “high.” FIPS 200 describes seventeen families of security controls. The level of security controls that a federal organization selects is a function of the identified impact levels for each of the three categories of security objectives. SP 800-171, in contrast, treats all impacts as “moderate” and addresses directly the single security objective of confidentiality. There may be practical benefits to this simplification, but it tends to “normalize” impact and “marginalize” the security objectives of system integrity and availability.
SP 800-171 claims a “well-defined” structure that consists of “a basic security requirements section” and a “derived security requirements section.”
In FIPS 200, the “Minimum Security Requirements” are very high-level, brief statements of objective, e.g.:
Access Control (AC): Organizations must limit information system access to authorized users, processes acting on behalf of authorized users, or devices (including other information systems) and to the types of transactions and functions that authorized users are permitted to exercise.
For the same family, “Access Control,” SP 800-53, revision 4, states 25 separate controls and several dozen available enhancements, and these are described over 51 pages of the publication. For this “AC” family, a total of 35 controls and enhancements must be met for the SP 800-53 “moderate” baseline.
The requirements for the “Access Control” family in SP 800-171 are much closer to FIPS 200 than to SP 800-53. They demand much less and are expressed at a much higher level. They consist of just two “Basic Security Requirements” and 22 “Derived Security Requirements;” all of the Derived Security Requirements are stated in a single sentence, and the whole treatment of the “AC” family of requirements in SP 800-171 takes less than one page.
The “Requirements” section of SP 800-171 informs companies about the safeguarding objectives. As concerning “Access Control,” for example:
3.1 ACCESS CONTROL
Basic Security Requirements:
3.1.1 Limit information system access to authorized users, processes acting on behalf of authorized users, or devices (including other information systems).
3.1.2 Limit information system access to the types of transactions and functions that authorized users are permitted to execute.
Derived Security Requirements:
3.1.3 Control the flow of CUI in accordance with approved authorizations.
3.1.4 Separate the duties of individuals to reduce the risk of malevolent activity without collusion.
[and so forth]
As the label implies, the “Derived Security Requirements” are based on the security controls in NIST SP 800-53, starting from the “moderate” controls baseline, but reduced to a subset through “tailoring” and expressed as brief principles rather than instructions. This was purposely done in order to avoid controls that are uniquely federal, unrelated to protecting just the confidentiality of CUI, or that NIST expects are routinely satisfied by nonfederal organizations without specification.
SP 800-171 lists 109 discrete requirements allocated to the 14 families that are to apply to CUI in nonfederal information systems. These requirements are “mapped” to identify relationships in the SP 800-53 security controls. The 109 requirements in SP 800-171 reference 85 controls from SP 800-53. This is considerably less than either the ordinary “moderate” baseline or the voluntary Framework, or even the SP 800-53 “low” baseline. Requirements described in SP 800-171 map to more SP 800-53 controls than DoD presently applies in its counterpart rule to protect UCTI:
References to NIST SP 800-53 Controls
Comparison of the number of cited controls should be undertaken with care. The intent of SP 800-171 is not to require contractors to comply strictly with controls (or control enhancements) from SP 800-53 just because NIST has provided tables that “map” the relationship. The intent is for companies who become subject to SP 800-171 to comply with the narrative statements in the Basic Security Requirements and the Derived Security Requirements for each of the 14 families; references to relevant controls or control enhancements from SP 800-53 are for information only. To satisfy the requirements of the SP 800-171, adoption of the SP 800-53 controls is not mandatory—or even encouraged. The strategy of NIST SP 800-171, instead, is to state performance or capability-based requirements that elaborate upon core principles drawn from FIPS 200 but do not express the “how to” rules of SP 800-53.
Through its 2013 regulations on UCTI, DoD has acted to impose limited security controls on its contractors that have access to unclassified technical information with military or space application. DoD’s UCTI regulations invoke 51 of the SP 800-53 cyber controls (61 if “enhancements” cited in the DFARS are counted).
Federal authorities are well aware of the need to examine carefully what constitutes a sufficient level of controls and will seek industry views on this subject. Further education is needed to be sure that the affected contractor universe understands that the SP 800-171 requirements relate to but are independent from the catalog of SP 800-53 controls and control enhancements. But the uncertainty or dimensions of special agency-dictated controls and upward tailoring will remain very much in the minds of industry.
Implications of NIST SP 800-171 for Industry
As the federal government moves to impose its version of cybersecurity rules on nonfederal information systems and service providers, private industry will raise many questions of need, relevance, suitability, efficiency, burden, cost and justification. While some companies in the federal supply chain undoubtedly lack appropriate or even rudimentary controls, many companies already will have measures in place and may be subject to different if not conflicting sources of obligation or oversight as to those measures. Achieving the positive purpose of protecting sensitive federal CUI in nondefense contractors must be affordable, accomplished without costs disproportionate to benefits, without wholesale exclusion of capable and trustworthy companies and without new barriers that separate federal agencies from technology innovation in the commercial marketplace.
In the pending version of SP 800-171, NIST in several important ways recognizes these concerns:
- The enumerated security requirements are tailored down significantly from the controls and enhancements of the SP 800-53 “moderate” baseline.
- NIST accepts that nonfederal organizations can implement a variety of potential security solutions, either directly or through the use of managed services to satisfy CUI security requirements.
- NIST emphasizes that the many additional controls described in 800-53 are “non-prescriptive”: While listed with the intent “to promote a better understanding of CUI security requirements,” 800-53 controls are “not intended to impose additional requirements on nonfederal organizations.”
57 NIST SP 800-171 (Final Public Draft), at 8 & n.19.
- SP 800-171 recognizes that nonfederal organizations have specific safeguarding measures in place to protect their information that also may be sufficient to satisfy the CUI security requirements.
- By mapping of NIST 800-171 requirements to other regimes, such as ISO/IEC, NIST appears to recognize that many companies already rely on other standards and practices to achieve the security sought for CUI. NIST also offers guidance on how to locate equivalent controls for 800-171 with the core functions of its voluntary Framework.
- NIST explicitly recognizes that companies may choose to create separate security domains to handle and protect CUI without increasing the organization’s “security posture” beyond what it needs for its core business or other operations.
- Because nonfederal organizations may lack the means to satisfy every CUI security requirement, NIST allows that they “may implement alternative, but equally effective, security measures to compensate for the inability to satisfy a particular requirement.”
58 Id. at 5.
Measures that protect federal information when in contractor hands also will protect valuable contractor information where the controls are employed across an organization. The federal government and its contractors share an interest in protecting the confidentiality of contractor intellectual property against extraction or theft, whether by criminal organizations, business organizations, state-sponsored hackers or state actors. But that does not mean that the federal government should impose its security rules upon private companies beyond the information system and service domains of a company where CUI (or UCTI) are resident or utilized.
Industry undoubtedly will be concerned about the prospect that federal agencies will take the requirements of SP 800-171 as the “floor” and use tailoring to layer additional obligations upon them. Industry’s concerns are natural and justified. As explained, NARA’s proposed CUI rule does not resolve the question of whether and how agencies will insist upon higher levels of information safeguarding. Nor does the present draft of SP 800-171 resolve the question. Under the heading, “The Requirements,” SP 800-171 states:
Additional CUI security requirements beyond those requirements described in this publication may be justified only when such requirements are based on federal law, regulation, or governmentwide policy and indicated in the CUI Registry as CUI specified. [sic] The provision of safeguarding requirements for CUI in a particular specified category will be addressed by NARA in its CUI guidance and in the CUI FAR; and reflected as specific requirements in contracts or other agreements.
This language—presented in a footnote—begs the question because it is unaccompanied by any of the implementation particulars that would make it more useful guidance. NARA has published the CUI Registry, but it far from complete. NARA’s proposed CUI rule expressly anticipates that agencies will determine additional security controls or dissemination constraints for categories or subcategories of CUI that are identified as “CUI Specified.” But the CUI Registry today contains no more than placeholders for this important content. Thus, even after NARA has published its proposed CUI rule, many important implementation details remain wholly conjectural.
The proposed SP 800-171 requirements will not have their intended effect, except perhaps as illustrations of potentially useful security practices, unless and until they are accompanied by both the final federal rule to define CUI and a new federal acquisition rule, of general application across all contracting, that will establish the methods and contract terms by which these requirements are imposed upon prospective federal contractors. Of this “triad,” the last two remain in gestation.
DHS Moves Out
This has not kept some federal agencies from moving out, nonetheless, with special measures to protect CUI that they consider of critical importance. DoD has made considerable progress with its UCTI regulations, first issued in November 2013. The Department of Homeland Security (DHS) recently issued a “class deviation” imposing a pervasive and highly demanding control regime on certain of its information when in the hands of its contractors.
http://tinyurl.com/lh59ywp.
The DHS Special Clause is to be used for existing as well as new “high risk” contracts where the contractor has access to “sensitive information” or its IT systems input, store, process or output such information. It is to be included in new solicitations, and DHS seeks to add the provision to existing contracts by bilateral modification.
The responsible DHS program manager is required to prepare a “Requirements Traceability Matrix” (RTM) when a contractor IT system is to be used with such sensitive information. That RTM is to be prepared in accordance with FIPS 199, meaning that security categorization will take into account the three objectives of confidentiality, integrity and availability as well as distinctions among “impact” levels. The RTM will generate the security controls that “must be implemented on the contractor’s IT system,” and these controls are set at “no less than ‘Moderate’ ” when a contractor’s IT system will be used with sensitive information that includes Personally Identifiable Information (PII), Sensitive Personally Identifiable Information (SPII) or Sensitive Security Information (SSI).
Where applied by contract, the clause obligates a contractor to follow multiple DHS-specific controls, policies and guidance. The contractor must receive an “Authority to Operate” and agree to and complete a “Security Authorization Process” which includes an independent third-party assessment. DHS insists upon a right to conduct “random periodic reviews” to ensure that the security requirements are met. Contractors subject to the Special Clause must afford broad audit access, and “continuous monitoring” requirements are imposed. Should there be a cyber event involving “known or suspected sensitive information,” the contractor must report “within one hour of discovery.”
On their face, these requirements are consistent with what might be expected to apply to “federal information systems,” and they appear to utilize processes (such as authorization to operate) drawn from the FedRAMP process that governs cloud security matters. However, the “class deviation” and the Special Clause do not appear to be limited just to companies who are under contract to DHS to operate federal information systems; rather, it seems to be the intent of DHS to apply these requirements to private contractors who have and use certain sensitive DHS information even if their access to or use of that information is through a nonfederal information system. If true, one can anticipate many industry objections because several of these requirements are onerous in part because they depart from customary norms even of private sector industry leaders. The requirements of DHS-specific authorization and DHS-directed third-party assessment, along with required audit access and monitoring, likely will generate objections as being unreasonably and unnecessarily costly and burdensome.
As illustrated by the new DHS initiative, important aspects of cyber supply chain requirements will be agency-specific. Agencies can tailor security controls to address the nature of information they protect and to reflect the risk of attack as well as the impact of loss of confidentiality. For low-risk situations involving information of relatively benign character, baseline controls could be tailored downward. Certainly, individual agencies have an interest in governing the reporting and response obligations that arise when a breach that affects the CUI of a particular agency occurs.
At the same time, if every agency imposes its own standard, and each agency applies its own oversight, the consequences could be impossibly disruptive and costly to many companies in the federal supply chain—especially to small businesses. Ultimately, though agencies will have the power to demand much of their suppliers, they cannot force companies to remain sellers in the federal marketplace. Some leading companies already refuse to sell directly to the federal government exactly because the unique federal compliance demands cannot be reconciled with their general, global business norms. In the end, the imposition of additional controls and risks will carry a price, and agencies will have to consider very carefully whether they can afford the price tag that would accompany imposition of unnecessary measures of prescription of NIST controls where other sufficient surrogates are in place.
Critical Missing Pieces
NARA is responsible to resolve what constitutes CUI that requires protection; its proposed rule recently entered the public comment period.
http://www.archives.gov/cui/.
SP 800-171 does not now differentiate among the sensitivity or significance of the various categories and subcategories of CUI. Nor do either the proposed CUI rule or SP 800-171 answer questions of what “tailoring” may be needed for CUI that agencies and NARA determine are to receive “CUI Specified” safeguards. All of these unknowns confront agencies, because those agencies will be uncertain whether their information is protected sufficiently, and the contractor community, which still does not know what costs and burdens to expect of heightened cybersecurity. These are very important considerations given the enormous volume and variety of CUI and the breadth of the potential application of federal cybersecurity controls to the commercial federal supply chain.
By definition, “nonfederal information systems” are those outside the boundaries of federal information systems. They may be systems of state and local governments, educational institutions, federal contractors and grantees, or those of other nongovernmental or even foreign organizations on which sensitive federal information (i.e., CUI) resides.
Thousands of nonfederal public and private sector enterprises host or use CUI, as it will be defined by the NARA CUI rule. They could become subject to the cybersecurity requirements of SP 800-171, and NARA is working (with the support of other federal agencies) on a FAR clause of general applicability that it intends will govern all federal agency purchases. The significance of this broad federal “enterprise level” initiative is potentially profound. As recognized by SP 800-171, the federal government relies upon nonfederal information systems and external information system service providers. When implemented by acquisition process and contract requirements, protecting that federal interest will impact the vast number of private companies that figure into the federal supply chain.
A Contractual Issue
That the federal government has important interests to protect, in the confidentiality of its CUI and UCTI, is inarguable. It is certain that other agencies will follow DoD’s lead in imposing at least “recommended requirements” upon the information systems of contractors that store, use or transmit this information. NIST will be a primary source for the security controls. But these requirements are not self-imposing upon federal contractors. Rather, federal agencies will utilize the means of “acquisition planning and contract administration” to achieve the intended protection of the confidentiality of CUI.
Acquisition planning implicates many possibilities. Civilian agencies might require offerors to meet at least the minimum cyber controls of SP 800-171 as a condition of eligibility for the award of contracts that involve the use, transmittal or generation of CUI. Federal civil agencies could consider the presence of minimally sufficient cyber controls, as suggested by SP 800-171, as necessary to demonstrate a contractor is “responsible” and therefore eligible for award.
http://tinyurl.com/lwjmq3j.
SP 800-171 focuses on systems to protect CUI. The present draft of SP 800-171 acknowledges its dependency upon the pending NARA CUI regulation to categorize, designate and mark CUI. Until the CUI rulemaking process is complete, SP 800-171 exists in a “vacuum” as to both what information is to be subject to controls and how industry is to be informed or self-determine whether information is subject to CUI controls. NARA has taken upon itself the task of producing a “single Federal Acquisition Regulations (FAR) clause that will apply the requirements of the proposed [CUI] rule to the contractor environment.”
Even without the “single FAR rule” that NARA contemplates, agencies can employ NIST SP 800-171 to assist in their “acquisition planning,” as a basis for contractually required CUI cyber safeguards, and for “contract administration” measures. An approach utilizing SP 800-171 would be a responsible and informed way for agencies to improve their assurance of contractor cyber protection without risking overreach or demanding cyber control practices at odds with even best commercial norms. SP 800-171 is intended to require contractors to safeguard CUI, but it recognizes both the existence and suitability of cyber control strategies and methods that do not require direct implementation of NIST’s SP 800-53 controls that were fashioned for federal information systems.
Until CUI is defined, a control regime is articulated and acquisition mechanisms are in place, neither the government nor industry will know what information is subject to the safeguarding requirements, whether or which cybersecurity requirements apply or what contractual duties (or liabilities) accompany the cybersecurity obligations.
Reporting Cyber Incidents
Absent from SP 800-171 are specific instructions for reporting of cyber incidents.
SP 800-171 makes little reference to NIST’s “Framework for Improving Critical Infrastructure Cybersecurity,” beyond offering guidance on how to use “mapping tables” to relate the controls required by SP 800-171 to counterparts in the five families of controls in the Framework. This seems odd because the Framework was developed in collaboration with industry to assist organizations, voluntarily, to adopt and apply risk-based measures to manage their cybersecurity risk. In the Framework, NIST observed that organizations “will continue to have unique risks—different threats, different vulnerabilities, different risk tolerances—and how they implement the practices in the Framework will vary.”
The same propositions hold true for the agencies whose CUI merits protection and for the private sector enterprises that may become subject to SP 800-171 controls. Similar risk-informed flexibility should guide the expectations and demands of individual federal agencies and their oversight. Informed forbearance from dictated controls, unnecessary oversight or administrative obligation or unreasonable demands will reduce compliance and implementation burdens on federal contractors.
Conclusion
The federal supply chain includes companies that are entrusted with federal information. The ICT systems of these companies are at constant risk of cyber attack. Considering the threat, and the national interest in protecting the many categories of sensitive federal information, it is necessary and proper for civilian federal agencies to use their authority over acquisition methods and contract requirements to improve cybersecurity and information assurance of nonfederal information systems. These measures should be taken only after the government is able to determine and designate the information to be protected, with due regard for the sensitivity of information and the consequences of its release or compromise, and with recognition of the diversity of companies affected and the presence of responsible choices among available cybersecurity controls.
Learn more about Bloomberg Law or Log In to keep reading:
See Breaking News in Context
Bloomberg Law provides trusted coverage of current events enhanced with legal analysis.
Already a subscriber?
Log in to keep reading or access research tools and resources.