Strict data consistency - Cacheonix guarantees that all servers in the cluster see the updates even when the servers fail or join the cluster. This allows applications to read the data from memory without going to the database periodically. Memcached doesn't provide data consistency. This means that applications using Memcached may see different data for the same keys which is generally unacceptable. Applications using Memcached must resort to expiring the cache periodically in order to get the most current data that leads to further performance degradation.
Superior performance - Cacheonix offers a coherent local cache so that an application using Cacheonix reads frequently data from the memory of the local JVM. Cacheonix partitions are co-located with the application, so parts of the data stored in Cacheonix is always read from memory which provides an additional performance boost. Application using Memcached must always go through the slow network to access the cached data.
Fastest-possible serialization protocol - Cacheonix is written in pure Java. Cacheonix offers developers an ability to implement Externalizable interface that provides fastest-possible wire-level serialization speed. Application using Memcached must read and write data using a heavy serialization protocol that is slower and takes significantly more bytes to transfer cached data that adds to the latency.