Access control in message-driven systems
summary
Message-driven systems enforce zero trust by generating a token per message that includes a payload hash claim and a message ID so consumers can verify signature, detect tampering and reject replays. Policies run in an Open Policy Agent PDP collocated as a PDP sidecar collocated with the application for low latency, and enforcement points call it to allow or deny actions. At high volumes use self-signed tokens or local auth sidecars; replay checks use the message ID replay check.
AI-written summary · about
chapters (20)
- 0:00Access control in messaging
- 1:09Why access control matters
- 2:16Zero trust foundation
- 5:02Zero trust principles
- 8:17Balancing zero trust costs
- 9:03Synchronous communication
- 12:15Asynchronous messaging identity
- 14:46Message types overview
- 16:54Commands queries events
- 17:16Enforcement points in flows
- 19:45Bounded contexts boundaries
- 21:55Open Policy Agent
- 25:33Zero trust implementation
- 26:54Token generation binding
- 28:49Verifying message tokens
- 31:02High-volume messaging
- 32:50Analytical workloads identity
- 33:31Event sourcing expiration
- 36:42User removal scenarios
- 38:48Summary advice
related talks
36mAccess control in message-driven systems
14mAPI access control strategies beyond JWT bearer tokens
25mOAuth Gets You In, Zero Trust Keeps You Safe
14mThe Problem Isn’t Access, It’s Actions
24mExpanding beyond authentication with emerging authorization standards
38mZero Trust Security for your APIs
17mZero Trust access with zero waiting, zero pain, and zero compromises
43mSecure your Modular Monolith by pretending it's a cat