Which key/value store is the most promising/stable?

Question

I\'m looking to start using a key/value store for some side projects (mostly as a learning experience), but so many have popped up in the recent past that I\'ve got no idea

Answer 1

As the others said, it depends always on your needs. I for example prefer whatever suits my applications best.

I first used memcached to have fast read/write access. As Java API I´ve used SpyMemcached, what comes with an very easy interface you can use for writing and reading data. Due to memory leaks (no more RAM) I was required to look for another solution, also I was not able scale right, just increase the memory for a single process seemed to be not an good achievement.

After some reviewing I saw couchbase, it comes with replication, clustering, auto-failover, and a community edition (MS Windows, MacOs, Linux). And the best thing for me was, the Java client of it implements also SpyMemcached, so I had almost nothing else to do as setup the server and use couchbase instead of memcached as datastore. Advantage? Sure, my data is now persistent, replicated, and indexed. It comes with a webconsole to write map reduce functions for document views in erlang.

It has Support for Python, Ruby, .Net and more, easy configuration through the webconsole and client-tools. It runs stable. With some tests I was able to write about 10k per second for 200 - 400 byte long records. Reading Performance was way higher though (both tested locally). Have a lot of fun making your decision.