In “Anatomy of Cassandra’s Read Operation”:
Key cache: Key cache stores a map of recently read partition keys to their SSTable offsets. This facilitates faster read access into SSTables stored on disk and improves the read performance but could slow down the writes, as we have to update the Key cache for every write.
The key cache is all about SSTable offsets. If the Memtable flush is asynchronous, then updating key cache does not slow down the write operation I guess? It may slow dow the flush operation, but client will never see that.