Caching: Speed at the Cost of Truth
Cache-aside, write-through, TTLs, invalidation and stampedes. Lesson 3 of Level 8: Data-Intensive Systems. Based on DDIA.
Level 8 · Data-Intensive Systems · Lesson 3 of 15
Cache-aside, write-through, TTLs, invalidation and stampedes.
Continue learning
- Indexes: Fast Reads, Slower Writes — Previous lesson
- Replication: Leaders, Followers & Failover — Next lesson
- Level 8: Data-Intensive Systems — Back to the level
Try it hands-on
- Cache: Hits, Misses, TTL & Invalidation — Serve product reads through a cache; tune TTL and size, update the database and watch stale reads and stampedes.
- Caching Lab — Hit rates, TTLs, invalidation, stampedes and CDNs.
All lessons in Data Systems
- Storage Engines: B-Trees vs LSM-Trees
- Indexes: Fast Reads, Slower Writes
- Caching: Speed at the Cost of Truth
- Replication: Leaders, Followers & Failover
- Replication Lag: When Your Own Write Disappears
- Partitioning: Hash, Range, Hotspots & Rebalancing
- Transactions & What ACID Actually Promises
- Isolation Levels: Dirty Reads to Serializable
- Lost Updates: Optimistic vs Pessimistic Locking
- The Trouble With Distributed Systems
- Consistency Models, CAP & Consensus
- Logs & Kafka: Partitions, Offsets and Ordering
- Consumer Groups & Delivery Semantics
- Batch vs Stream Processing & Data Pipelines
- Event Sourcing & CQRS