Policy Evaluation Logic and Precedence
You'll learn how cloud providers combine multiple IAM policies and determine whether to allow or deny a request.
The Deny-By-Default Foundation
Cloud IAM starts from a position of implicit deny: unless explicitly granted, every action is blocked. This zero-trust baseline means your policies must affirmatively permit actions. If no policy statement matches your request, access is automatically denied.
Think of it like a secure building where every door is locked by default — you need the right key (permission) to open any door.
The Explicit Deny Trumps All Rule
When evaluating whether to allow an action, the IAM engine follows a strict precedence hierarchy:
- Explicit Deny — If any policy contains an explicit "Deny" statement matching the request, access is immediately blocked. Period. No other policy can override this.
- Explicit Allow — If at least one policy explicitly allows the action and no explicit denies exist, access is granted.
- Implicit Deny — If no policies address the action at all, access is denied by default.
Policy Evaluation Order
The IAM service evaluates all applicable policies — identity-based policies attached to the user, group policies, role policies, and resource-based policies. It doesn't stop at the first match; it examines every relevant policy to check for explicit denies.
This means even if you're granted broad permissions through a managed policy, a single explicit deny in an inline policy or permission boundary will block that action.
Analogy: Imagine collecting votes where a single "no" vote (explicit deny) vetoes the entire decision, regardless of how many "yes" votes (allows) you receive.
Key Takeaway: Policy evaluation follows a strict precedence: explicit denies always win, then explicit allows are checked, and finally implicit deny serves as the default — always evaluate all policies before making an access decision.