Securing Distributed Cache: Achieving Secure-by-Default with Key Challenges &...
summary
Netflix runs an in‑house distributed cache serving 400 million requests per second across 200 clusters and 22,000 servers holding 14.3 petabytes of data; the system uses mutual TLS, certificate lifecycle signing, per‑connection auth via an authorization agent, and a Java thick client with non‑blocking IO. A config drift causes workers to call the policy store on every request, consuming >50% CPU; debug uses CPU profiling and heap dumps, fixes roll out, and auth adds ~5% cost.
AI-written summary · about
chapters (5)
- 0:00Securing Netflix's cache
- 5:27EV cache client security
- 21:05EV cache server auth
- 31:18Incident lessons learned
- 37:22Security trade-offs
related talks
47mCloud Security at Scale and What it Means for Your Application
27mBehind the Scenes: Running Critical Services @ Netflix
36mHow Netflix Is Solving Authorization Across Their Cloud [I]
19mKeynote: Making and Keeping Netflix Highly Available
15mFrom Chaos to Clarity: Deciphering Cache Inconsistencies in a Distributed...
54mDistributed Caching in Real-Time Payment Systems
43mCaching half a billion user transactions
21mHandling Traffic Spikes Through Automated Prioritized Load Shedding