A Hitchhiker’s Guide to Caching Patterns
summary
Caching is a trade‑off between staleness and speed and the talk shows concrete patterns and a working demo: a Python Flask + SQLAlchemy app, Hazelcast server, a Java map‑loader for read‑through, a Hazelcast write‑behind with a 20s delay window that coalesces updates, and a Hazelcast Jet pipeline doing CDC to keep the cache evergreen as cache‑ahead. Demos prove cache‑aside reduces DB load, read/write‑through hides the DB, write‑behind boosts throughput but risks lost updates unless replicated.
AI-written summary · about
chapters (15)
- 0:00Introduction to caching trade-offs
- 3:15Why not build your own cache
- 9:59Basic cache usage patterns
- 14:23Demo: Cache aside implementation
- 17:39Write operations with caching
- 18:42Read-through caching explained
- 20:05Demo: Read-through with Hazelcast
- 24:02Write-through caching explained
- 25:49Demo: Write-through with Hazelcast
- 28:45Write-behind caching pattern
- 32:37Refresh-ahead vs cache-ahead
- 34:45Cache-ahead with change data capture
- 37:38Demo: Cache-ahead implementation
- 42:39Recap of caching patterns
- 45:06Conclusion and resources
related talks
43mWhere Is My Cache? Architectural Patterns for Caching ...
30mCaching patterns for lazy devs for lazy loading
46mCaching Made Bootiful
47mCaching beyond simple put and gets
44mCatch Me If You Cache: Cache Patterns and Pitfalls in Production
27mCaching for Cash
42mCaching for Cash
25mArchitectural Caching Patterns for Kubernetes