# SCIM

SCIM is an open standard for provisioning and managing user and group identities across independent systems through a common HTTP API and identity schema.

Cloud concept. Updated 2026-11-15. Page: https://blowstack.com/tools/cloud-concepts-hub/scim

> System for Cross-domain Identity Management (SCIM) automates the creation, update, and deactivation of accounts as people join, move within, or leave an organization. It complements sign-in federation by keeping application identity records synchronized with an authoritative identity source, reducing manual administration and orphaned accounts.

Practice exams covering SCIM: [SAP-C03 AWS Certified Solutions Architect - Professional](https://blowstack.com/practice-exams/aws-certified-solutions-architect-professional).

## 6 key aspects of SCIM

SCIM standardizes identity lifecycle synchronization while leaving authentication, application authorization, and session revocation to complementary controls.

1. **Standard identity schemas and APIs**: SCIM defines common representations for users and groups and an HTTP-based management protocol. This reduces application-specific integration work, while schema extensions accommodate additional identity attributes.
2. **Automated identity lifecycle management**: Provisioning clients create, update, deactivate, or delete target accounts according to source-directory changes and application assignments. Validate each target's behavior because disabling an account and deleting it are not equivalent.
3. **Separation from authentication**: SCIM manages identity records rather than authenticating users or issuing sign-in assertions. Pair it with SAML or OpenID Connect when an application needs both automated provisioning and federated sign-in.
4. **Group synchronization and authorization boundaries**: SCIM can synchronize group membership that applications use for access decisions. Application roles, cloud permissions, and resource assignments remain separately configured authorization controls.
5. **Identity correlation and attribute mapping**: Attribute mappings translate source-directory fields into target schemas. Stable correlation identifiers and tested matching rules prevent account duplication or accidental reassignment when names or email addresses change.
6. **Secure, observable synchronization**: Provisioning requires protected endpoints, securely managed credentials, and monitoring for failed or delayed updates. Measure deprovisioning latency and handle session or token revocation separately where required.

## What is SCIM

System for Cross-domain Identity Management (SCIM) standardizes how systems exchange and maintain identity records. SCIM 2.0 defines JSON-based resource schemas and an HTTP protocol, primarily specified in RFC 7643 and RFC 7644. Its core resources are users and groups, with attributes such as usernames, display names, email addresses, account status, and group membership. Schema extensions allow applications to represent additional information, such as employee numbers and departments.

A typical cloud deployment has a provisioning client in an identity platform and a SCIM service provider endpoint in a target application. The client reads identities from an authoritative directory, applies assignment rules and attribute mappings, and sends provisioning requests to the target. For example, assigning a new engineer to an application can create a user record and add that user to an engineering group before the engineer first signs in.

SCIM is not an authentication or single sign-on protocol. SAML 2.0 and OpenID Connect establish authenticated sessions, whereas SCIM manages the accounts and memberships applications use. A common architecture combines SAML sign-in with SCIM provisioning. Just-in-time provisioning during sign-in is different: it can create an account when someone authenticates, but does not by itself provide a complete mechanism for updating or deactivating accounts when that person stops signing in.

SCIM exposes resources through endpoints such as /Users and /Groups. The protocol defines operations for creating, querying, replacing, modifying, and deleting resources, although supported capabilities differ between implementations. PATCH enables partial updates when supported. Filtering and pagination help clients locate and reconcile records. The target assigns each resource an id; an externalId can correlate that record with an identifier managed by the provisioning client. Choose stable matching attributes carefully because email addresses and usernames can change.

Lifecycle behavior requires explicit design. A departing employee might be represented by setting the user's active attribute to false rather than deleting the record, preserving audit history and application-owned data. Deactivation, deletion, and group removal can have different effects in different targets. Provisioning is generally asynchronous, so a successful change in the source directory does not imply immediate enforcement everywhere. Existing application sessions, API tokens, and previously issued cloud credentials might remain usable until separately revoked or expired.

SCIM transports identity state, not a universal authorization policy. An application can map provisioned groups to roles, but its own permission model still determines access. In AWS IAM Identity Center, for example, an external identity provider can provision users and groups through SCIM, while permission sets and account assignments separately determine AWS account access. Provisioning a group alone does not grant it an AWS role.

Operationally, protect SCIM endpoints with HTTPS and appropriately scoped credentials, commonly bearer tokens. Store credentials securely, rotate them before expiration, and monitor failures, throttling, and provisioning lag. Test attribute mappings, duplicate handling, unsupported group structures, and deprovisioning behavior before broad rollout. Standardization reduces custom integration work, but implementations still vary in supported attributes, filtering, PATCH behavior, and deletion semantics. Use reconciliation, retries with backoff, and audit logs to detect partial failures rather than assuming every lifecycle change has reached every application.
