Is write-behind ever safe for important data?

Only if you add durability the cache itself does not provide, like writing to a replicated log such as Kafka before acknowledging, at which point you have rebuilt a different architecture. As a plain cache pattern, write-behind trades a window of data loss for speed, so reserve it for data where losing a few seconds of writes costs effectively nothing.

Why not just use write-through everywhere and stop worrying about staleness?

Two costs. Every write pays a second round trip, which matters on hot write paths, and every written key consumes cache memory whether or not anyone reads it, which evicts genuinely hot data on write-heavy workloads. Write-through is the right tool for read-after-write data, and wasteful for data written often and read rarely.

What does cache invalidation actually mean in this context?

Removing or expiring a cached entry because the underlying data changed, so readers stop seeing the old value. The main mechanisms are TTL expiry, explicit deletes triggered by writes, and versioned keys that make old entries unreachable. Most production setups layer at least two of these, using explicit deletes for freshness with a TTL as the safety net.

Cachingchapter 4 of 8 · 5 lessons

Cache Write Strategies

How and when writes propagate from cache to the database.

← All chapters View in cheatsheet

Your cache is now answering most reads. Then somebody edits a price.

For a moment there are two answers to the same question, and only one of them is true. What you do in that moment is the whole of this chapter. Write to both places and every write pays for both. Write to the cache and flush later, and a crash takes the difference with it. Skip the cache entirely and readers keep the old value until something removes it. Three answers to one question: when a write happens, does the copy get updated now, later, or never?

Lessons

5 in this chapter

Write-ThroughEvery write updates the cache and the database together, so reads never see a gap.2 min
Write-BehindAcknowledge from the cache, flush to the database later, and accept the crash risk.2 min
Write-AroundWrites skip the cache entirely, and only reads decide what deserves memory.3 min
Cache ConsistencyBetween the write landing and the copy going, somebody is reading the old value. The question is who, and for how long.2 min
Picking a CombinationRead and write strategies are chosen per data type, not once per system.3 min