Authentication and Authorization Decisions
Key Takeaways
Authentication verifies a claimant through an authenticator. Authorization decides whether a principal can perform a specific action.
Carry the subject, resource, action, and relevant policy context through an authorization test.
An authorized operation still requires execution and observable target evidence.
For an API, verify token validity and endpoint permissions separately.
Follow one attempted operation
Authentication verifies the claimant is entitled to act as a particular identity or client. A verifier checks evidence such as a password or a cryptographic challenge response. The result can support establishing a session. Authorization evaluates permission for an operation on a resource. The request can be from a person or an application. Record which principal is acting, what action is requested, the exact target resource, and any context required by the access policy.
For example, a payroll analyst signs in through a trusted identity provider. That verifies the principal for the application. Payroll may allow that analyst to read payroll summaries and deny editing employee payment destinations. The same successful authentication supports both attempts. The authorization outcome changes with the operation. Changing the analyst’s sign-in method would not give the person the missing business permission.
| Observed evidence | What it establishes |
|---|---|
| Valid identity-provider sign-in and verified application session | Successful authentication for the session |
| A policy permits the user to view their own region’s report | Authorization for that action, resource and context |
| The same user tries editing a report with no edit grant | Evaluate the separate edit operation under its policy |
| Connector accepted a write and the changed field is present on the native account | Evidence of execution for that target |
| A bearer token has passed the required validity checks for an API request | Check endpoint permission, scope and resource separately |
Evaluate the inputs to policy
A role-based design ties permissions to granted roles. A policy can also evaluate attributes of the subject, object, operation and environment. For example, the rule for viewing a payroll summary could require the reader to have the appropriate role and the summary to belong to the reader’s region. In a test, fix the same signed-in identity and the same view action, then try a summary from a different region. This changes the resource context while holding the authentication result and action fixed. A different test holds the resource fixed and changes the operation from read to edit. A third test changes the principal to an authorized editor so the system demonstrates that the exact update action works for its intended audience.
Platform administration illustrates another boundary. In SailPoint, a tenant user level grants permissions within the platform. An external payroll entitlement is permission on the target payroll system. Someone who can run certification campaigns may have the permissions required to govern payroll access within SailPoint. Whether they can edit the payroll system depends on that application’s own authorization. A reviewer assignment in a campaign also addresses a different purpose from being allowed to act as a platform administrator.
Diagnose a token or permission fault
For an application programming interface (API), inspect which authorization server issued the access token, whether it is valid for the target, and how you are presenting it. Check the endpoint’s authorization requirements after establishing a valid token. The Hypertext Transfer Protocol (HTTP) status 401 indicates that valid authentication credentials were missing for that request. A 403 indicates that the server understood the request but refuses it. SailPoint includes insufficient rights and lack of required user context among documented examples. Inspect the response details so an error label does not become the only diagnosis.
Suppose a monitoring script uses a valid token to read a permitted resource and gets 403 from a write endpoint. Check whether that write is even within the script’s purpose. If authorized by the business, verify the required endpoint scopes, subject context and privileges under the product docs. A fresh token issued on the same terms can leave the write forbidden. Conversely, a script with the right roles and an expired access token needs valid credentials before its requested privileges even matter.
For a diagnostic test, record one success and one expected denial on a separate controlled resource or action. Verify the subject and resource in both cases, then confirm that an expected denial left the target unchanged. Capture any successful native effect separately from successful sign-in, token issuance, or a configured grant.
See NIST’s digital identity guidelines and attribute-based access decisions, SailPoint’s authorization boundary and platform user levels, and the HTTP status semantics.
An analyst signs in with the corporate identity provider and can read payroll reports. Editing a report is denied by the application’s access policy. Which condition changed between the attempts?
The authentication factor category
The issuer’s signing certificate
The authorization decision for the edit action
The native account’s unique identifier
A valid script token can read an API resource, but the write endpoint returns 403. What should the engineer check before changing credentials?
Whether the script should perform the write, and the endpoint’s required scope, user context, and permissions
Whether the login button can be moved to a new URL
Whether changing the access token’s expiration time grants the write
Whether the account aggregation schedule can substitute for endpoint authorization
Sections you finish are checked off in the contents.