Skip to content

DOMAIN 5 · Identity and Access Management

Digital Identity: Proofing, Authentication, and Federation

Distinguish IAL, AAL, and FAL, and choose identity controls according to the consequences of failure.

A digital identity represents a subject within a particular service. It does not automatically establish that person’s legal identity, employment, or entitlement to a resource. NIST SP 800-63-4 separates identity proofing, authentication, and federation so that organizations can address different failures with appropriate controls. This article uses the July 2025 revision; an older reference to SP 800-63-3 can describe different requirements.1

For CISSP study, connect each mechanism to a business decision: who needs access, what could happen if the wrong person receives it, and what evidence supports the required confidence.

Dimension Decision supported Example failure
Identity Assurance Level, IAL Confidence in the claimed real-world identity Enrolling an impostor as a customer
Authentication Assurance Level, AAL Confidence that the claimant controls the account’s authenticators Accepting a stolen password
Federation Assurance Level, FAL Confidence in identity information conveyed between identity provider and relying party Accepting a forged or injected assertion

These are separate dimensions, not three names for one security score. Their selected levels can differ. A service can need strong protection for a pseudonymous account without collecting extensive identity documents. Conversely, careful identity proofing does not prevent a later account takeover.1

Authentication factors and phishing resistance

Section titled “Authentication factors and phishing resistance”

NIST’s AAL2 requires proof of control of two distinct authentication factors. Two passwords remain the same kind of factor. AAL3 requires a public-key cryptographic authenticator with a non-exportable private key and phishing resistance, with proof of possession of that key plus either an activation factor or a password.1

Evaluate the complete implementation rather than assigning an assurance level from a product name. A deployed authenticator, enrollment process, session, and recovery route must all support the intended protection. A convenient fallback can otherwise become the easiest route into a strongly protected account.

In a federated service, an identity provider authenticates a user and supplies information that a relying party uses. The relying party still owns decisions about access to its resources. Outsourcing the login screen does not outsource authorization policy.

NIST’s FAL model addresses protection of the federation transaction, including forged assertions and, at higher levels, additional attacks such as assertion injection. Selecting FAL requires understanding which party issues assertions, which service consumes them, and the impact of an identity-provider compromise.1

Worked example: an employee benefits service

Section titled “Worked example: an employee benefits service”

Consider a fictional employer with an internal lunch-ordering site and a benefits portal containing sensitive personal records. The two services share an identity provider, but their consequences of failure differ.

For the benefits portal, the design team documents how employee identity is established, which authenticators are accepted, how assertions are checked, and how account recovery works. It also decides which benefits records an authenticated employee may see. A successful login as employee A must never permit reading employee B’s records. This example illustrates the distinction between identity confidence and resource permission; it is not a NIST-prescribed assurance assignment.

  • “A high IAL makes every login strong.” Proofing and subsequent authentication solve different problems.
  • “MFA means phishing-resistant.” Factor count alone does not establish resistance to a fraudulent verifier.
  • “SSO grants access to every connected application.” Federation conveys identity information; application policy determines permission.
  • “More identity data is always safer.” Unnecessary collection adds privacy and handling risks. NIST asks organizations to consider privacy and customer experience alongside security.1

Continue with access control to translate identity information into permissions. Control assessment explains how to gather evidence that the intended identity controls actually operate.

  1. NIST, SP 800-63-4, Digital Identity Guidelines, July 2025; Introduction, Digital Identity Model, and Digital Identity Risk Management assurance-level selection sections. ↩ ↩2 ↩3 ↩4 ↩5

Independent, unofficial study notes. Not endorsed or reviewed by ISC2. CISSP is a registered trademark of ISC2.

Editorial policy & attribution