---
title: "Discussion on \"How AWS Actually Evaluates IAM Policies (Identity vs Resource)\" | Hashnode"
description: "Discussion on \"How AWS Actually Evaluates IAM Policies (Identity vs Resource)\". How AWS evaluates IAM policies: implicit deny, explicit deny priority, and the union of identity-based and resource-based policies in the same account"
url: "https://hashnode.com/posts/how-aws-actually-evaluates-iam-policies-identity-vs-resource/6aa6ab4a28c6fb8953d4bf60"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [How AWS Actually Evaluates IAM Policies (Identity vs Resource)](https://eadeshina.hashnode.dev/how-aws-actually-evaluates-iam-policies-identity-vs-resource?utm_source=hashnode&utm_medium=feed)

How AWS evaluates IAM policies: implicit deny, explicit deny priority, and the union of identity-based and resource-based policies in the same account

Sep 13 · 7 min read · eadeshina.hashnode.dev  
Updated Sep 14  
Read the article

-   [#aws-iam](https://hashnode.com/tag/aws-iam)
-   [#policy-evaluation](https://hashnode.com/tag/policy-evaluation)
-   [#identity-security](https://hashnode.com/tag/identity-security)
-   [#cloud-security](https://hashnode.com/tag/cloud-security)

## Written by

[Adeshina Emmanuel](https://hashnode.com/@emmanueladeshina01)

@emmanueladeshina01 · Lagos, Nigeria

  
[View profile](https://hashnode.com/@emmanueladeshina01)

AI Infrastructure Security Engineer | Cloud, Kubernetes & AI Systems Security | Inference, Observability & Platform Security

I’m Emmanuel Adeshina, an infrastructure security engineer focused on cloud, Kubernetes, distributed systems, and AI infrastructure. I’m interested in how complex systems handle identity, authorization, networking, runtime security, reliability, and failure across distributed environments. I’m building \*\*RELIASTRA\*\*, an independent external dependency observation system for engineers. It observes third-party APIs and infrastructure from outside the systems being measured, records real observations, detects persistent failures, correlates incidents, and produces verifiable evidence. The broader goal is dependency intelligence: turning independent observations of external infrastructure into a durable trust layer for software systems. I write and research around cloud security, AI infrastructure, Kubernetes, distributed systems, reliability, and measurement integrity. \*\*I build infrastructure, investigate how it fails, and produce evidence for what actually happened.\*\*

## Responses

Markdown supported · ⌘↵ to post

No responses yet.