Apache Cassandra has two mechanisms that can be used to cache table data: the key cache, and the row cache. As the name suggests, the key cache simply caches the location of the keys in RAM, providing a boost in efficiency to disk seeks for rows as well (Gilmore 2011). Similarly, enabling row-caching stores the entire row in cache. Hits on the row cache perform significantly better than reads from disk, as reads from RAM happen in a matter of nanoseconds (Dean, Norvig 2012).
Row-caching requires the size to be set in cassandra.yaml as well as the number of rows (per partition) to be set at the table level. Do note that row-caching is disabled by default, as its use is known to interfere with a node's ability to serve requests. ...
