There are enough NoSQL databases out there to make quite the kettle of alphabet soup, but Couchbase is making some moves that may help it float to the top of the pot.
The company will announce the general availability of Couchbase 4.0 at an event this week in New York. Important new features include NIQL, a SQL-like query language for running analytics, says Constellation Research vice president and principal analyst Doug Henschen. While there are plenty of these types of languages available in the NoSQL domain, "they've got a good start here," he adds.
"They're also focusing on multi-dimensional scaling, which is about being able to independently scale the resources you need for data, query and indexing workloads," Henschen says. "You're hearing about that a lot on big data platforms. That's key because it's about balancing scale and speed requirements without breaking the bank. If everything has to go together, you might needlessly scale up network, storage or compute capacity when only one or two dimensions might do the trick."
Couchbase has also added JSON document handling in the last year. "They're trying to give you the best of both worlds: Document data handling and huge scalability," Henschen says.
No Magic Bullet
Couchbase has managed to snare some extremly high-profile, data-rich customers, including PayPal and LinkedIn. Still, it's a crowded market with many alternatives to Couchbase.
"There's a world of products out there," such as MongoDB, which is popular with developers but has been addressing scalability issues, Henschen says. "To appeal to developers is one thing," he adds. "IT has to worry about the long term."
Couchbase and MongoDB have been engaging in a tit-for-tat PR battle over the question of scalability, firing claims and counter-claims. That said, Couchbase's sweet spot seems to be in between NoSQL platforms such as Riak and Cassandra, which are geared toward extremely high-scale, global deployments and MongoDB, which is often deployed with one to three nodes, Henschen says.
The Bottom Line
Ultimately, there's no single correct NoSQL platform for all use cases, he adds.
"There's a lot of nuance in that world of choices," Henschen says. "Reads versus writes, for example, are two different things. And there are application use cases where MongoDB is just fine at high scale."
