Hacker News new | comments | show | ask | jobs | submit login

Cassandra, but mainly because we have a couple devs who are excellent at debugging Java/JVM issues. Cassandra is young enough you really need to expect to get your hands dirty in its code to keep things humming. (I may be more forgiving of Cassandra because, unlike Mongo, Cassandra's major version number is 0).

Another team seems awfully happy with Hadoop, although they've been bitten by an HDFS related bug. I can't speak too much to their experience.

Memcached of course is wonderful, and the little bit I've used Redis it has lived up to its promises.

Riak is the only non-relational database I can think of that I've only heard positive things about, but I've never personally used it.

I'm a big fan of Redis precisely because it has a short list of promises and executes them well.

So many of these other new data stores promise "infinite scale" and leave out the "as long as it fits in RAM" part.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact