It’s not uncommon to look at open-source projects as the communes of IT hippies, where they spend their days growing organic code they’re more than willing to share with their neighbors, and their nights sitting around the campfire singing “Kumbaya” and dreaming of a world without proprietary code or software licenses. It all sounds so great at the time. Ultimately, though, some members grow up, take jobs with The Man and realize something’s missing from the FOSS visions of their youth — namely, that it’s difficult to spread the gospel if it requires a lot of work for would-be devotees.
So, up spring commercial software vendors that continue paying homage to open source, but that sell out, if you will, by adding new proprietary features, licenses and/or paid support to make some money by courting business customers. And then they turn on each other: “Use this Linux distro, not that one,” “MySQL in the cloud is great, but our service is far superior to theirs.” It’s not necessarily a bad outcome, just the reality of growing up and realizing that business sometimes is a zero-sum game. Strangely, though, the NoSQL community doesn’t appear to be falling into this trap — for now.
The reason for such unity among the NoSQL set seems pretty clear. It’s generally accepted that every NoSQL database has its own benefits ideal for specific use cases, so there’s no reason to squabble among each other as to which one is the best. The strategy makes sense on the surface, as it’s theoretically better to keep the overall NoSQL buzz strong and let interested organizations learn about everything and choose the product(s) best suited for their particular needs — likely as a complement to an existing relational database. As I’ve been told on numerous occasions by NoSQL executives, it’s not about one NoSQL product versus another, or even about NoSQL versus relational databases, but about choosing the right tool for any given job. I wrote about Objectivity’s foray into the NoSQL community after 20-plus years selling its object database last week, and Chief Architect Darren Wood described just how welcoming everybody has been.
As an example, I recently came across the NoSQL Tapes, a web site launched in January, ran by Cloudant employee Tim Anglade,
funded whose summer-long NoSQL expedition (before he began at Cloudant) was supported by scale-out storage startup Scality and Objectivity, and containing dozens of videos of NoSQL bigwigs from just about every project and company discussing their wares technologies. Not only is Objectivity present on the NoSQL Tapes site talking about its InfiniteGraph graph database, but so are representatives talking about Cloudant (CouchDB), CouchOne (CouchDB) and Membase (Membase Server) (now a joint unity called Couchbase), HBase, Dynamo, MapReduce and Scality. In the queue are videos featuring Cloudera (Hadoop), DataStax (Cassandra), Voldemort, Basho (Riak), MarkLogic, Acunu and a whole lot more. One NoSQL vendor is literally paying to host paid to help launch a site now ran by an employee of another and featuring advertisements, essentially, for even more potentially competitive products.
Below the surface, though, cracks are showing in NoSQL utopia. In December, 10gen Founder and CEO Dwight Merriman told my colleague Stacey Higginbotham that most companies will only choose a single NoSQL database, and he firmly believes many will choose MongoDB, the document database his company has commercialized. James Phillips echoed that sentiment to me when explaining that the merger of Membase and CloudOne to create a Membase-CouchDB hybrid called Couchbase will give NoSQL adopters a holistic database that saves them having to look elsewhere. I would argue, too, that Cassandra proprietor DataStax (formerly Riptano) would rather companies use its products instead of opting for HBase, as Facebook, which created Cassandra, keeps doing. These stances make perfect sense, too, because businesses exist to make money, and that’s more difficult to do when you have direct competition or when you tell prospective customers that your product will only solve part of their problem.
The challenge for NoSQL members going forward — if they care to — will be to keep the community spirit strong as commercial interests grow even stronger. Given the relative immaturity of the market, there’s certainly something to be said about maintaining the communal vibe and openly discussing their various projects in mass meetups in Silicon Valley and across the country. A rising tide lifts all boats, after all.
But it seems almost inevitable that something — a major uptick in business, intra-market consolidation or acquisitions by large vendors — will come along to upset the good feelings. The NoSQL community will still get together, but it will be as exhibitors and speakers at industry conferences, and when they talk about each other’s products, it might be largely derogatory, despite the fact that they all were created to solve largely the same problem. Think Red Hat (s rht) and Novell, or Oracle (s orcl) and just about every other relational database vendor. Whether it’s for better or for worse, the possibility of such a jaded future makes it worth embracing NoSQL’s brotherly love for as long as it lasts.
Image courtesy of Flickr user Rojer.
Related content from GigaOM Pro (sub req’d):
- Report: NoSQL Databases — Providing Extreme Flexibility and Scale
- With Scalable Data Stores Around, Is NoSQL a Non-Starter?
- The Incredible, Growing Commercial Hadoop Market