I have a bit of a hangover from two back-to-back NoSQL conferences, MongoDBWorld and Couchbase Connect. So I thought I'd try a hair-of-the-dog cure and write about yet another document database: Clusterpoint. This one is "as a service" -- i.e., a cloud-based solution -- and launched in the U.S. market just a few days ago.
Clusterpoint is a Latvian company that's been around since 2006. Its first product was an on-premises search engine, which has evolved into a more standard document database, now located in the cloud.
I asked Clusterpoint's CEO, Zigmars Rasscevskis, why you'd want to use his solution as opposed to, say, MongoDB or Couchbase. In fact, Rasscevskis sees other DBaaS (database as a service) solutions as his main competiton -- Cloudant, DocumentDB, DynamoDB, and the like.
Clusterpoint offers ACID-compliant, multi-document transactions today, a feature MongoDB only has on its roadmap. Clusterpoint also incorporates the features of a full-text search engine. Rasscevskis sees this as an opportunity to use one database instead of three (a search engine, an RDBMS, and a document database).
Interestingly, you can also run Clusterpoint on-premises if you've signed up for an account. Obviously, this is helpful when you're dealing with high data volumes or need to reduce latency -- or if you feel your data is too sensitive to put in the cloud, even though Clusterpoint offers data-at-rest encryption.
The hook is that you can start on the Clusterpoint platform for free and move to paid offerings as your data needs grow. That said, it's fair to question whether the market will stomach a very specific proprietary database offering from a single "as a service" provider. If you code to the Clusterpoint database for free, you have no real option to leave later without the expense of recoding to another.
Also, there's the question of where you place your bets. It's one thing if you decide to tie your fate to the likes of Amazon. Sure, you'll be forced to pay the piper if Amazon arbitrarily raises its rates, because the cloud is a lock-in game, but Amazon will certainly be around for a long time. Smaller independent cloud providers don't inspire the same level of confidence.
For a while now we've seen a glut of key-value store startups, from Aerospike to Voldemort. Now it's happening on the document database front. Progress!