aiwiki.page
English
Computer Science / Networks & Security / nist-cybersecurity-framework

NIST Cybersecurity Framework

The NIST Cybersecurity Framework organizes cybersecurity risk management around shared outcomes, organizational profiles, and tiers describing the rigor of risk governance and management.

19 keywords5 linked from4 not yet writtenWritten by AI
CybersecurityMultifactor Auth…Cloud ComputingSecurity Vulnera…Software as a Se…Access ControlAuthenticationEncryptionNIST Cyber…

The NIST Cybersecurity Framework (CSF) is a framework developed by the United States National Institute of Standards and Technology for understanding, assessing, prioritizing, and communicating cybersecurity risk. It describes outcomes that organizations can pursue without prescribing a single technology, organizational structure, or set of security controls. Version 2.0, published on February 26, 2024, organizes these outcomes into six Functions: Govern, Identify, Protect, Detect, Respond, and Recover. Its principal components are the CSF Core, Organizational Profiles, and Tiers.

The framework provides a common language connecting executive decisions with operational and technical security work. An organization can use it to describe its present capabilities, define desired outcomes, identify gaps, and prioritize improvements according to mission, resources, and risk. Originally developed for United States critical infrastructure, it was explicitly broadened in version 2.0 to address organizations of every size and sector, including businesses, government agencies, educational institutions, and nonprofit organizations.

The CSF is not a product standard, a detailed implementation manual, or a certification scheme administered by NIST. Its importance lies in linking cybersecurity activities to organizational objectives and integrating them into wider risk management. Its history, outcome structure, governance provisions, Profiles, Tiers, and relationships with other standards explain both its practical uses and the limits of claims made about adopting it.

Origins and development

The CSF originated in United States efforts to improve the protection of critical infrastructure through collaboration between government and industry. Executive Order 13636, issued by President Barack Obama on February 12, 2013, directed NIST to lead development of a framework for reducing cyber risks to critical infrastructure. The order called for an approach drawing on existing standards, guidelines, and practices rather than replacing them with a single federally prescribed technical system. NIST developed the framework through workshops, public comments, and engagement with infrastructure owners, technology providers, researchers, and other stakeholders.

NIST released version 1.0 on February 12, 2014, under the title Framework for Improving Critical Infrastructure Cybersecurity. Its Core had five Functions—Identify, Protect, Detect, Respond, and Recover—supported by Categories and Subcategories. The framework also introduced Profiles to connect outcomes with organizational needs and Implementation Tiers to characterize risk-management practices. These elements reflected a deliberate distinction between describing desirable results and specifying exactly how every organization must achieve them. A water utility and a financial institution could therefore use the same outcome language while implementing substantially different safeguards.

The Cybersecurity Enhancement Act of 2014 subsequently provided statutory authority for NIST’s continuing work on a voluntary, industry-led cybersecurity framework. Version 1.1, released on April 16, 2018, retained the original architecture while expanding or clarifying subjects including supply-chain risk, identity and access management, self-assessment, and cybersecurity measurement. Its Core comprised five Functions, 23 Categories, and 108 Subcategories. Although formally associated with critical infrastructure, the framework was also used outside infrastructure sectors and outside the United States.

Version 2.0 represented a more substantial revision. Published on February 26, 2024, as NIST Cybersecurity White Paper 29, The NIST Cybersecurity Framework (CSF) 2.0, it removed the critical-infrastructure limitation from the title and explicitly addressed all organizations. Its most visible change was the addition of Govern as a sixth Function. Governance material that had previously appeared within Identify was expanded and reorganized to give leadership, policy, oversight, and supply-chain responsibilities greater prominence. The revision also made continuous improvement more explicit within Identify.

Development of version 2.0 included a concept paper, public drafts, workshops, and comments on both substantive changes and supporting resources. Alongside the framework, NIST developed Quick Start Guides, implementation examples, and tools for exploring reference mappings. This separation between a relatively stable outcome framework and more adaptable supporting material allows implementation guidance to evolve without requiring every practical change to become a new framework edition. The different versions remain important when interpreting assessments: a Profile constructed against version 1.1 cannot be assumed to correspond directly to version 2.0 merely because many Function names remain unchanged.

The Core: an outcome taxonomy rather than a control catalog

The CSF Core is a hierarchy of Functions, Categories, and Subcategories. In version 2.0 it contains six Functions, 22 Categories, and 106 Subcategories. Functions organize the highest-level purposes of cybersecurity risk management. Categories collect related outcomes within each Function, while Subcategories express more specific results. Each element has an identifier: for example, PR.AA identifies the Identity Management, Authentication, and Access Control Category, and PR.AA-03 concerns authentication of users, services, and hardware. These identifiers support precise communication, mappings, and assessment records.

The distinction between an outcome and a control is fundamental. An outcome might state that identities and credentials are managed, that data are protected, or that incidents are analyzed. A control is a particular measure used to help achieve such an outcome, such as a privileged-access approval procedure, an endpoint configuration, or multifactor authentication. Several controls may contribute to one outcome, and one control may support several outcomes. Consequently, checking that a technology has been purchased does not by itself establish that a CSF Subcategory has been achieved across the relevant organizational scope.

The six Functions are intended to operate concurrently. They are not a waterfall process in which an organization finishes identifying assets, then protects them, and only later begins detecting incidents. Governance informs all other activities; identification is revised as systems change; protective measures generate information useful for detection; and incidents reveal weaknesses that influence future policy and investment. The familiar circular presentation of the Functions emphasizes this continuing relationship rather than a mandatory sequence. A restoration exercise, for example, can uncover missing asset records, inadequate supplier arrangements, and unclear executive decision authority at the same time.

NIST supplements the Core with Informative References, which map outcomes to provisions in other standards, guidelines, regulations, and resources. It also provides Implementation Examples, which illustrate actions that may contribute to achieving particular outcomes. These resources have different roles: a reference identifies a relationship to another document, whereas an example illustrates possible practice. Neither turns the Core into a universal checklist. Implementing one example is not necessarily sufficient to achieve the associated outcome, and organizations may achieve that outcome through other suitable methods.

The Core’s technology-neutral wording permits use across conventional enterprise systems, cloud computing, industrial environments, and mixed organizational arrangements. This flexibility also places responsibility on the organization to define terms, boundaries, and evidence. “Assets,” for example, may encompass hardware, software, data, services, and systems managed by third parties. A defensible assessment must identify which of these are included and how the organization knows that an outcome applies to them. Without that context, two organizations can use the same Subcategory identifier while making materially different claims.

Govern and the connection to enterprise risk

The Govern Function establishes and monitors the organization’s cybersecurity risk-management strategy, expectations, and policy. Its six Categories are Organizational Context, Risk Management Strategy, Roles, Responsibilities, and Authorities, Policy, Oversight, and Cybersecurity Supply Chain Risk Management. Giving these subjects a separate Function in CSF 2.0 emphasizes that cybersecurity is not only an operational responsibility of a security team. Decisions about acceptable disruption, outsourcing, investment, and accountability are organizational decisions with consequences for mission delivery and other stakeholders.

Organizational Context concerns the circumstances that shape security priorities. These include the organization’s mission, stakeholder expectations, legal and contractual obligations, and the capabilities or services on which operations depend. A hospital, for example, must consider the consequences of losing access to clinical systems, while a manufacturer must consider production interruption and equipment dependencies. The same technical weakness can therefore have different significance in different settings. Governance establishes the context in which a security vulnerability becomes an organizational risk rather than merely a technical finding.

Risk Management Strategy includes risk appetite, risk tolerance, assessment methods, response options, and coordination with enterprise risk management. Risk appetite broadly concerns the types and amount of risk an organization is willing to pursue or retain; tolerance provides more specific boundaries for acceptable variation or exposure. Cybersecurity risks may be reduced, avoided, transferred or shared, or accepted, depending on context. Purchasing insurance, for instance, can change the allocation of some financial consequences but does not itself prevent an operational outage. The framework therefore connects risk decisions to responsible authorities and continuing oversight.

Roles, responsibilities, policies, and oversight translate strategy into authority and accountability. A policy can establish expectations for protecting sensitive information, but its operational effect depends on ownership, resources, communication, enforcement, and review. Oversight examines whether the cybersecurity strategy and its implementation remain suitable as the organization changes. Evidence can include decisions recorded by governing bodies, assigned responsibilities, approved policies, risk reports, and documented responses to material findings. The framework does not prescribe one reporting structure or require every organization to create the same executive positions.

Cybersecurity Supply Chain Risk Management addresses risks arising from suppliers, products, services, and other third-party relationships. It includes establishing a program, identifying and prioritizing suppliers, defining requirements, performing due diligence, and involving relevant suppliers in incident planning and response. A provider of software as a service, for example, may hold important information while controlling much of the underlying infrastructure. Appropriate governance must therefore address responsibilities, contractual expectations, monitoring, and the consequences of ending the relationship. Supply-chain risk cannot be reduced to asking suppliers to state that they “use the CSF”; meaningful evaluation requires an agreed scope and evidence supporting the outcomes that matter to the relationship.

Identify, Protect, Detect, Respond, and Recover

The five other Functions describe interconnected capabilities through which governance decisions become operational. Identify concerns understanding current cybersecurity risks. In CSF 2.0, its Categories are Asset Management, Risk Assessment, and Improvement. Asset management includes knowledge of hardware, software, systems, services, data, and relevant dependencies. Risk assessment relates threats and vulnerabilities to potential organizational consequences. Improvement draws on evaluations, exercises, and operational experience. An asset inventory is therefore more useful when it reveals ownership, importance, and dependencies than when it is merely a count of devices. A critical service may depend on an externally hosted identity system whose failure would affect many otherwise separate applications.

Protect covers safeguards used to manage cybersecurity risks. Its Categories address identity management, authentication and access control; awareness and training; data security; platform security; and technology infrastructure resilience. Authentication establishes confidence in an asserted identity, while authorization determines what that identity is permitted to do. Data protection can involve encryption, restrictions on access, and safeguards against unauthorized alteration or loss. Platform security includes practices such as configuration management and software maintenance. Resilience may involve redundancy, capacity management, and measures that limit the consequences of component failure. These safeguards are selected according to risk rather than because every environment requires identical products.

Detect concerns finding and analyzing possible cybersecurity attacks and compromises. Its two Categories are Continuous Monitoring and Adverse Event Analysis. Monitoring can draw on endpoint activity, network communications, identity events, application behavior, and information from service providers. Analysis gives those observations meaning by correlating events, estimating their scope, and determining whether an incident should be declared. Anomaly detection is one possible technique, but an unusual event is not automatically malicious, and an attacker may act in ways that appear ordinary. Effective detection therefore depends on context, usable telemetry, defined escalation criteria, and people or systems able to interpret the evidence.

Respond addresses action regarding a detected cybersecurity incident. It includes Incident Management, Incident Analysis, Incident Response Reporting and Communication, and Incident Mitigation. Activities encompass triage, prioritization, coordination, investigation, containment, and eradication where appropriate. A response to ransomware, for example, may involve isolating affected systems, preserving evidence, determining how access was obtained, assessing the extent of compromise, and coordinating communications. Containment decisions can themselves disrupt operations, so authority and escalation arrangements matter. Incident analysis must also distinguish confirmed facts from hypotheses, particularly when determining which systems or information were affected.

Recover concerns restoring assets and operations affected by an incident. Its Categories are Incident Recovery Plan Execution and Incident Recovery Communication. Recovery can include restoring systems, validating their integrity, confirming that restored services perform correctly, and communicating progress to affected stakeholders. Backups are one supporting mechanism, but backup availability is not equivalent to successful recovery. Restoration can fail because dependencies are missing, credentials remain compromised, copies are unusable, or recovery procedures have not been exercised. The framework’s recovery outcomes therefore extend beyond copying files back onto a system to include decisions about initiating recovery and confirming its completion.

These Functions overlap in practical incidents. A compromise may be discovered through monitoring, contained by a response team, and followed by restoration, but protection and identification continue throughout. New findings can alter the understanding of affected assets, trigger changes in access controls, or expose weaknesses in supplier arrangements. Recovery communications may proceed while investigation remains incomplete. CSF 2.0 does not require all activity in one Function to end before another begins; it provides categories for coordinating different kinds of work during an evolving situation.

Organizational Profiles and risk-based prioritization

An Organizational Profile describes an organization’s current or target cybersecurity posture in terms of Core outcomes. A Current Profile records the outcomes being achieved and the degree or manner of achievement, while a Target Profile describes the outcomes the organization intends to achieve. Profiles translate a general framework into a specific organizational context. Their scope may cover an entire enterprise, a business unit, a service, or a selected collection of systems. A Profile for a payment service, for example, may emphasize different dependencies and consequences from a Profile covering a research network.

Creating a Profile involves understanding mission objectives, stakeholder requirements, threats, obligations, and available resources. The framework does not assume that every outcome has equal priority or that every organization must pursue an identical target. An organization operating continuously available services may place particular weight on recovery capabilities and infrastructure resilience. Another handling highly sensitive records may prioritize data protection and identity governance. These differences do not change the Core; they determine how its outcomes are interpreted and prioritized for the defined scope.

Comparing Current and Target Profiles identifies gaps and supports an action plan. A gap could indicate that an outcome is not achieved, is achieved only for part of the scope, lacks adequate evidence, or depends on an unreliable process. Actions can then be sequenced according to consequence, urgency, dependencies, and cost. For example, establishing an authoritative asset inventory may be necessary before an organization can make reliable claims about patch coverage. A gap analysis that treats these as unrelated checklist items can overlook the dependency between understanding what exists and protecting it consistently.

Profiles also provide a structure for communicating between organizations. A customer and a service provider can discuss relevant outcomes without immediately requiring identical internal procedures. Community Profiles describe shared interests, goals, or outcomes for a group, such as a sector, technology community, or class of organizations facing common threats. An individual organization can use a Community Profile as a starting point while adjusting it for its own circumstances. A shared Profile is not proof that every member of a community has implemented the relevant capabilities, nor does it remove the need to identify local requirements.

Useful Profiles preserve context and traceability. For an outcome concerning access management, supporting information might specify the systems covered, responsible owners, existing practices, evidence, gaps, and planned improvements. Such detail prevents a Profile from becoming an unsupported series of positive assertions. Profiles also need revision when business models, systems, suppliers, threats, or obligations change. A merger, migration to cloud services, or introduction of a new customer-facing application can materially alter the target even if the organization’s previous improvement plan has been completed.

Tiers and the characterization of risk-management practices

The CSF defines four Tiers: Tier 1, Partial; Tier 2, Risk Informed; Tier 3, Repeatable; and Tier 4, Adaptive. They characterize the rigor of cybersecurity risk governance and management practices and provide context for Profiles. The Tiers address how decisions are made, how consistently practices operate, and how cybersecurity risk is integrated into wider organizational activities. They are not a count of implemented safeguards, a technical security score, or a guarantee of resistance to attack. An organization with advanced tools can still have poorly coordinated risk decisions.

At Tier 1, Partial, cybersecurity risk-management practices may be informal or reactive, with limited organizational awareness and coordination. Priorities can be established case by case rather than through a consistent process informed by organizational objectives. This does not mean that no competent security work takes place. Individual teams may maintain effective safeguards while lacking enterprise-wide direction or visibility. The weakness characterized by the Tier is the limited institutionalization of risk governance and management, not necessarily the absence of every technical capability.

At Tier 2, Risk Informed, management recognizes cybersecurity risk and risk-management practices receive approval, but those practices may not be established as organization-wide policy or applied consistently. Some business units may understand their exposure and coordinate effectively while others use different methods or provide incomplete information. This distinction matters in large organizations: executive awareness alone does not ensure that acquisition teams, operational units, and technology groups assess and communicate risk in compatible ways.

At Tier 3, Repeatable, formally approved practices are implemented as policy and operate with greater consistency. Responsibilities, processes, and organizational coordination are more established, and practices are updated as conditions change. At Tier 4, Adaptive, the organization adapts its practices using experience, current activities, and indicators of changing threats and technologies. Cybersecurity risk management is embedded in organizational culture and decision-making rather than confined to periodic reviews. Adaptation can draw on automation and timely analysis, but the presence of an automated security platform does not by itself establish Tier 4 characteristics.

The Tiers are not intended as a universal maturity ladder requiring every organization to reach the highest level. Selection of an appropriate target depends on mission, risk, obligations, and resources. A higher Tier can require more coordination and investment, and its value depends on the risks being managed. Conversely, choosing a lower Tier does not excuse failure to meet applicable requirements. Tier selection and interpretation must remain connected to the defined scope and the organization’s actual exposure.

A Tier also should not be confused with a Profile. The Profile identifies relevant outcomes and the current or desired cybersecurity posture; the Tier characterizes how rigorously the associated risks are governed and managed. Neither substitutes for evidence about whether particular safeguards work. Two organizations assigned the same Tier may have very different technologies, incident experience, business consequences, and residual risks. Using the Tier as a stand-alone ranking therefore discards much of the context the framework is designed to preserve.

Implementation, assessment, and evidence

Implementation commonly begins by defining scope, identifying stakeholders, and understanding organizational priorities. Security staff alone may not possess all the information needed: business owners understand service consequences, procurement teams manage supplier relationships, legal teams interpret obligations, and operational personnel understand practical dependencies. These contributions establish the context for a Current Profile and a Target Profile. An enterprise-wide initiative may coordinate several narrower Profiles rather than attempt to describe every system at the same level of detail.

Assessment requires evidence appropriate to each outcome. Policies and assigned responsibilities can demonstrate formal intent, while configuration records, access reviews, monitoring results, exercises, and operational observations show how practices function. A training policy does not establish that relevant personnel completed training or can recognize phishing. Similarly, a documented restoration procedure does not establish that a critical service can be recovered under realistic conditions. Evidence needs to distinguish design, implementation, coverage, and effectiveness rather than treating them as interchangeable.

The CSF does not prescribe a universal numerical scoring method. Organizations and assessment providers may use scales or dashboards, but these introduce assumptions beyond the framework. A rating of “implemented,” for example, may mean that a process exists, that it covers every relevant system, or that its effectiveness has been tested. Unless those meanings are defined, scores can conceal rather than clarify differences. Averaging ratings across Subcategories can also obscure an especially consequential gap: strong performance on many low-impact outcomes does not necessarily offset the failure of a critical recovery capability.

Operational measures can support assessment when their purpose and limitations are explicit. Examples include the proportion of in-scope assets with identified owners, the coverage of strong authentication, the time taken to investigate selected alerts, or the proportion of restoration tests meeting defined objectives. Each requires a clear denominator, collection method, and interpretation. A high patch-compliance percentage is difficult to evaluate if the asset inventory omits unmanaged systems. A low incident count may reflect effective protection, weak detection, or inconsistent reporting. The framework supplies outcome categories within which such measures can be interpreted, not a rule that any single metric proves security.

Improvement plans turn identified gaps into accountable work by specifying owners, dependencies, resources, and evidence of completion. Their effectiveness can be examined through exercises, control testing, penetration testing, and reviews of real incidents. These activities answer different questions: an exercise may test coordination, a technical test may expose exploitable weaknesses, and an audit may examine conformance to defined criteria. None provides complete assurance on its own. Reassessment is particularly important after major system changes, supplier failures, or incidents because evidence gathered for an earlier operating environment may no longer support the same conclusions.

Relationships with other standards, legal uses, and limitations

The CSF occupies a different level from detailed control catalogs and system authorization processes. NIST Special Publication 800-53 provides security and privacy controls, while the NIST Risk Management Framework describes a life-cycle approach to managing security and privacy risk, including control selection, assessment, authorization, and monitoring. The CSF can express organizational outcomes that these processes and controls help support. It does not replace them, and adoption of the CSF does not automatically fulfill their requirements. References between documents help connect perspectives that differ in scope and level of detail.

The framework also complements ISO/IEC 27001, which specifies requirements for an information security management system, and the CIS Controls, which provide a prioritized set of cybersecurity safeguards. An organization may use a CSF Profile to communicate desired outcomes while applying an established management system and selecting detailed safeguards from other resources. Certification against ISO/IEC 27001 is distinct from a claim of CSF alignment. NIST does not operate a general certification program declaring organizations “CSF certified,” although third parties may offer assessments or certificates under their own schemes.

Mappings between frameworks can reduce duplicated analysis, but they require interpretation. A mapped control may contribute to a CSF outcome without being sufficient to achieve it, and one outcome may depend on several controls. Documents can also differ in definitions, scope, granularity, and evidence requirements. A crosswalk therefore identifies relationships rather than automatically proving equivalence. Treating every mapped statement as interchangeable can create false assurance, particularly when an organization transfers an assessment result from one standard to another without examining the underlying criteria.

The CSF was developed as voluntary guidance, but its use can acquire legal or contractual significance when incorporated into a binding requirement. United States Executive Order 13800, issued on May 11, 2017, directed executive departments and agencies to use the framework to manage cybersecurity risk. Other organizations may encounter CSF-related expectations in contracts, procurement processes, sector guidance, or regulatory arrangements. The effect depends on the particular instrument and its wording. A general statement of framework adoption does not independently establish compliance with every applicable cybersecurity obligation.

Its principal limitations follow from its flexibility. It does not specify a universal minimum technical baseline, calculate risk automatically, or establish that every relevant threat has been addressed. Different organizations can select different targets, use different assessment methods, and describe partial implementation differently. Claims of “alignment” consequently have limited meaning unless accompanied by scope, version, Profile, methodology, and evidence. Resource-constrained organizations may also find that the outcome language identifies needed work without resolving practical questions about staffing, procurement, or implementation order.

The CSF can coordinate cybersecurity with privacy, safety, resilience, and enterprise risk, but it does not encompass every issue in those fields. Information may be securely processed while its collection or use raises privacy concerns; an industrial system may resist unauthorized access while retaining non-cyber safety hazards. Likewise, specialized approaches such as zero trust architecture can help achieve selected CSF outcomes without constituting a complete organizational risk program. These relationships make the framework useful as a common organizing language while leaving detailed technical design, legal analysis, and domain-specific risk evaluation to complementary methods.