Databases, like ScyllaDB, should bypass the Linux page cache due to its inefficiency in serving database workloads. ScyllaDB uses its own highly efficient row-based integrated internal cache instead.
•10m read time• From thenewstack.io
Table of contents
The Problem with the Linux Page CacheOur Goal: Better Performance (and Control)Deeper into the ScyllaDB Cache DesignWhy Not Use a Buffer Cache?Using an Object Cache InsteadSummary1 Comment
Sort: