IAM Boundaries and Permission Guardrails
What you'll learn: How to use permission boundaries to cap the maximum permissions that identity policies can grant, preventing privilege escalation.
The Core Problem
Even with least privilege policies, there's a risk: what if someone with permission to create or modify IAM policies accidentally (or maliciously) grants themselves or others excessive permissions? You need a safety net that says "no matter what policies get attached, you can't exceed THIS ceiling."
What Are Permission Boundaries?
A permission boundary is a maximum permission policy you attach to an IAM entity (user or role). It acts as a guardrail—the entity can only perform actions that appear in BOTH:
- Their identity-based policies (what they're explicitly granted)
- Their permission boundary (their maximum allowed ceiling)
Think of it like a fence around a property. You might give someone keys to multiple buildings (identity policies), but they can only access buildings inside the fenced area (permission boundary). Even if they get a key to a building outside the fence, they still can't reach it.
How Evaluation Works
The effective permissions are the intersection:
- Identity policy says: "Can do A, B, C"
- Permission boundary says: "Can do B, C, D"
- Result: Can only do B, C
If the boundary doesn't include an action, it's denied—even if an identity policy grants it.
Common Use Cases
- Developer self-service: Let developers create roles, but only with permissions within a safe boundary
- Delegated administration: Allow team leads to manage IAM without risking full admin escalation
- Compliance enforcement: Ensure no one can bypass security controls, regardless of policy changes
Key Takeaway: Permission boundaries set an immutable maximum permission ceiling, ensuring that even those who manage IAM cannot escalate beyond defined guardrails—critical for delegation and defense-in-depth.