February 10, 2011

My Notes on the NoSQL Smashup

As with many developers, I love to tinker with new technologies.  Much like Dustin, I'd long admired CouchDB for it's novel approach, the fact that it matches the way we think about data and its simplicity.  It was hard to contain the excitement of bringing these things together.

In one particularly challenging situation, I'd recently had the opportunity to talk with the Silicon Valley Cloud Computing Group about the category of key value stores.  In particular, I was talking about the unique approach we'd taken with Membase to being a CP datastore.  The irony is that this was one day after a session on Document stores, where my new colleague J. Chris Anderson had spoken about CouchDB and CouchOne!

 At the time, I couldn't fill Sebastian or Shashank about the fact that we were going to kick the whole thing into high gear.  

 It's hard to overstate the possibilities here.  Consider for a moment…

 CouchDB and CouchOne have, over the last five-plus years and growing to millions of installs, been steadily gaining traction.  The refreshing (relaxing?) approach with CouchDB just makes sense to many developers like myself.

 Memcached, and increasingly Membase, are at the core of some of the largest sites out there.  With Membase Server we have eased the on-ramp, made it easy to scale up and down, and increased platform coverage such that we have lots of new deployments at sites like TheKnot and NaviNet.

 From an engineering standpoint, we have just assembled an amazing group of people.  Within days of setting about it, there was plenty of collaboration at the code level between the two groups.  Note that code commit from Dale Harvey of CouchOne improving Membase code was on February 4.  That's a full four days before we announced the merger and a mere 5 days after he'd come over to the U.S. to meet the rest of the team.

 This should be fun!