Solving Tricky Coordination Problems in Stateless .NET Services
summary
Claims a practical fix for cache stampedes by combining memo lock semantics with SET NX PX and Redis Pub/Sub, implemented as a small C# and Go library that interoperates with Python or serverless workers. Only one instance acquires the lease and computes the expensive value; other concurrent requests subscribe and receive an immediate notification when the value is ready. Demo shows one of two simultaneous requests does the work, the other waits and then returns the same cached result; supports lock renewal and timeouts.
AI-written summary · about
chapters (13)
- 0:00Intro
- 1:15Overview
- 3:25The Basics
- 8:57Documentation
- 9:38Sessions
- 10:21Coordination Problems
- 11:12Interface
- 14:49PubSub
- 22:42Cache Miss
- 26:29Locking
- 29:12Fencing
- 35:02Memo Lock
- 45:48More Intelligence
related talks
53mLeveraging Caching For Cloud Native Applications
35mBuilding Scalable Stateful Services
47mArchitectural Patterns for Caching Microservices
43mIntro to Stateful Services
54mDistributed Caching in Real-Time Payment Systems
14mMuCache: A General Framework for Caching in Microservice Graphs
36mCaching Strategies for an Erlang Based Web Stack
32mCache me if you can: Boosted application performance with Redis and client-side caching