MemcacheDB persistent implementation using LevelDB

Memcached is a data caching system which well known as in-memory database that provide faster data access compare with on-disk database. Persistency support in Memcached can be built by unification between Memcached as in-memory database and BerkeleyDB as the persistency backend store. However, MemcacheDB had lower performance compare to Memcached, due to the persistency overhead. This paper explored how to improving MemcachedDB performance by replacing BerkeleyDB with LevelDB, which had better performance than BerkeleyDB. The results on this research showed that the use LevelDB on MemcacheDB produced more capable operations than BerkeleyDB for either random or sequential data.