April 2011 Summaries

Cassandra's built-in key and row caches provide efficient data caching, allowing users to remove dedicated caching tools like memcached from their stack. With proper tuning, hit rates of 85% or better are possible with Cassandra. Key caching is enabled by default, while row caching provides the most dramatic gains in efficiency when feasible. Both caches can be configured and tested incrementally to optimize performance. Monitoring tools like jconsole GUI and DataStax OpsCenter Performance tab help track cache metrics exposed through JMX.