Looking at SRE Needs and Trends over Two Decades with a...
summary
Chubby starts as a leader election service for GFS and becomes a general-purpose small-key storage used for service discovery, configs and ACLs. Implementation uses Paxos consensus across five replicas, a watch/notification API, and client libraries; later additions include a read proxy and cache, per-client rate limiting and auth plumbing. Iterations succeed (cache, proxies) but a parallel dual-write deployment adds latency and operational toil and is retired after about a year.
AI-written summary · about
chapters (15)
- 0:00Introduction to Lock Service
- 1:37Lock Service original purpose
- 4:06Lock Service for GFS
- 7:11Chubby evolution
- 10:42Lock Service challenges adoption
- 12:29Lock Service cattle not pets
- 15:59Lock Service accountability isolation
- 18:30Lock Service API design Hyrum's Law
- 19:52Third lesson first mover disadvantage in infrastructure
- 22:22Fourth lesson the value of iterative improvements
- 24:26Lock Service failed iteration
- 26:55Lock Service file system?
- 27:57Lock Service 2018 challenges
- 31:07Lock Service lessons future
- 33:58Acknowledgments
related talks
58mSRE Your gRPC...
48mSRE Your gRPC—Building Reliable Distributed Systems Illustrated with gRPC
28mThe Curse of SRE Autonomy and How to Manage It
35mKeeping Ceph RGW Object Storage Consistent
49mWhat SRE Could Be: Systems Reliability Engineering
20mFine Grained Replicated State Machines for a Cluster Storage System
35mHow SRE Makes Electric Vehicles
38mThe SRE’s Crystal Ball: Predicting System Performance with...