Coordinated Access Control with @policy
summary
Booking.com turns multi‑team access disputes into a schema contract using the @policy directive, with domain teams authoring rules, Legal/Privacy and Identity adding cross‑cutting policies, and the router as the single-enforcement-point. Enforcement calls a central policy engine at plan time, strips denied fields so sensitive values never leave their owning service, and records every decision for audit. Adoption starts slow (about 10%) so an automated scanner and MR bot finds ~1,000 potential sensitive fields and drives 100+ teams to act, cutting onboarding from weeks to hours.
AI-written summary · about
chapters (9)
- 0:00Introduction to policy
- 1:09Data access challenges
- 6:02Solution as coordination problem
- 7:36Policy enforcement architecture
- 10:29Implementation benefits
- 12:00Adoption hurdles
- 13:23Automated adoption agent
- 15:08Key learnings
- 17:00Q&A
related talks
30mRedefining Access Control: Scaling Policy as Code for Humans and AI Agents
20mBuilding Scalable Access Control Frameworks in React
31mAccess Policies for Hybrid Cloud Environments
43mApplication Code of Conduct - Full-Stack Policy as Code
27mSecurity-by-construction: How to weave authorization into modern app stacks using Open Policy Agent
25mModernizing Authorization: From Basic Roles to Decoupled ABAC
52mIAM policy power hour
35mHow to fight misconfiguration and bad intentions with Kubernetes admission policy