Distributed cache

How this question breaks down

1. 1PatternsCore20 minWhere the cache sits relative to the database, who fills it, and what happens to it on a write.
2. 2EvictionCore29 minWhen memory is full, which key goes. Recency, frequency and plain FIFO, the approximations real engines run, and why one-hit wonders matter more than any of them.
3. 3ShardingCore26 minSplit the keyspace across many cache nodes so every client finds the same node for a key, and adding or losing a node costs a sliver of the hit ratio instead of all of it.
4. 4StampedesCore27 minStop one popular key's expiry, a mass expiry, or a flood of lookups for keys that do not exist from sending the whole read load to the database at once.

Work through the parts in order and you’ve covered the full design. Each part is one topic.