2.2 was exactly an "optimization" release, to bring what we had at a better level of maturity. Basically we'll try hard to don't add things to the API in the n

2.2 was exactly an "optimization" release, to bring what we had at a better level of maturity. Basically we'll try hard to don't add things to the API in the next releases, but just to open to new use cases changing the "backend" part, with cluster support for large fault tolerant deployment, and with diskstore for "bigdata". However there are a few important new things in Redis 2.2 from the point of view of the features, I think the main ones are: * non blocking replication, so that now slaves are able to serve data even when trying to resync with the master. * Check and Set with WATCH. * Write operations against keys with an expire set. * LRU eviction of keys in 'maxmemory' mode. * Support for SETBIT/GETBIT/SETRANGE/GETRANGE, basically this turn the string data type into a random access array.