Architectural Patterns for Caching Microservices
summary
Compares four concrete caching topologies and when each reduces latency or ops: in-process libraries (Guava, EHCache) give lowest latency but no eviction or cross-replica coherence; embedded distributed uses Hazelcast to form JVM clusters automatically; client‑server uses Redis/Memcached or Hazelcast clients and supports multi‑language apps; cloud cache uses a Hazelcast cloud discovery token for VPC‑peered low‑latency clusters; Kubernetes sidecar and reverse-proxy sidecar caching (NGINX/Envoy prototype) let apps stay unchanged but make cache invalidation the hard problem.
AI-written summary · about
chapters (18)
- 0:00Introduction to caching
- 1:03Why use caching in microservices
- 2:53Embedded cache pattern
- 7:54Embedded distributed cache demo
- 12:08Pros and cons of embedded cache
- 13:31Client-server cache pattern
- 17:06Client-server cache implementation
- 18:31Cloud cache pattern
- 21:00Cloud cache demo
- 26:03Pros and cons of client-server and cloud cache
- 27:29Sidecar cache pattern
- 30:58Sidecar cache implementation
- 32:00Pros and cons of sidecar cache
- 33:02Reverse proxy cache pattern
- 36:17Reverse proxy sidecar cache
- 41:03Reverse proxy cache challenges
- 44:37Decision tree for caching patterns
- 47:07Resources and closing
related talks
45mWhere is my cache? Architectural patterns for caching microservices by example
34mAdvanced Caching Patterns used by 2000 microservices
14mMuCache: A General Framework for Caching in Microservice Graphs
42m3 easy performance improvements in your microservices architecture
35mCaching at Netflix: The Hidden Microservice
32mCaching for Microservices - Introduction to Pivotal Cloud Cache - Pulkit Chandra
53mData Access and Data Sharing in Microservices
43mData Access and Data Sharing in Microservices