Maybe it's just me being naive, but facebook has got to be one of the most awesome places that there is to work.<p>I swear that every couple of months I hear about a situation that basically comes down to this:<p>Boss: "Hello, Ryan, we would like to do $impossible_thing""<p>Ryan: "Huh, that sounds kindof impossible. But I guess we could get a team of the smartest people in our industry together and build an entirely new suite of tools to make $impossible_thing into $totally_normal_thing..."<p>Boss: "Yeah, do it!"<p>6 months later<p>Ryan and boss (to internet): "Hey, guys, check out this thing we made. Here, you can have it for free. Have a nice day!"
I suppose fb caches the recent messages in memory, and use HBASE as archive storage most of the time.<p>It depends on the developer's favor to pick up a system. If it's me, I will pickup cassandra. The consistency model is not so hard to work with.
heh. I attended Jonathan Gray's talk at Hadoop World this year (which was pretty good!) and he talked a lot about Facebook's infrastructure. Gray has been a primary committer on the HBase project for a while (since inception?) and I remember people somewhat teasing that HBase was kind of the pet project of the NoSQL world. I've had fun hacking on it off and on myself. Gray amusingly said that HBase was being prepped for production in a huge way, but he couldn't give any details.<p>Well, I guess we know now! Definitely looking forward to all the future developments of HBase. There are a lot of interesting contribs being made.