Identity management
As you scale your use of the AWS Cloud, you need robust identity and permission management processes to help ensure that you follow the standard security advice of granting least privilege, or granting only the permissions required to perform a task. Robust identity management helps ensure that the right systems and people have access to the right resources under the right conditions. This also needs to be done while not overburdening operations capabilities with too many, too granular or too complex permission or identity constructs. The M&G Guide includes the recommendations of the Security Pillar for managing identities and access permissions across all cloud resources in your multi-account strategy in order to be migration ready, scale ready, and operating efficiently.
The Security Pillar describes the difference between human and machine identities. It also shows that centralized administration of human identities and access to your environments with an identity provider is a critical strategy to managing authentication and authorization across your enterprise. This is important for managing and governing, as it makes it easier to manage access across multiple applications and services because you are creating, managing, and revoking access from a single location. For example, if someone joins or leaves your organization, you can add or revoke that individual’s access for all applications and services (including AWS) from one location. This aligns with ITIL best practices, and reduces the need for multiple credentials and provides an opportunity to integrate with existing human resources (HR) processes. In AWS, we consider machine identities distinctly from human identities. Machine identities (like service roles) still reside within AWS IAM and are designed to uphold the principle of least privilege, but are not managed by your identity provider.
Define access policies and mechanisms that include granting least privilege access, sharing resources securely, and reducing permissions continually (including the removal of unused permissions) with AWS IAM Access Analyzer. Review how permissions are actually being authored, validated, and used over time, so that you can remove unnecessary permissions in accordance with the principle of least privilege. This would include adding observability rules for “last accessed” data, such as a timestamp depicting when an identity policy or principal (such as a user or role) last used a service or performed an action from supported services. This enables you to more easily identify unused permissions and improve your security posture by removing the permissions that are not necessary for the user, group, or role to perform a specific task. Both AWS and AWS Partners provide tools for the creation, review, and revoking of permissions in an automated manner throughout your software development lifecycle (SDLC) or development, security and operations (DevSecOps) cycles.