The NoSQL market has done anything but sit still since last year’s NoSQL Now conference. Last year, the big story was the expanding definition of NoSQL, which now includes graph databases and time-series databases, as well as so-called NewSQL databases. This year, however, the focus has shifted to highlight two topics more relevant to the enterprise: governance and transactions.

(10gen, the company behind MongoDB, changed its name to MongoDB Inc. shortly after the NoSQL Now conference.)

Jason Bloomberg, president of analysis firm ZapThink and author of “The Agile Architecture Revolution,” spoke about the topic of governance at the conference. “We need a way to manage policies for ongoing Big Data challenges. It’s not about dealing with datasets of a certain size, it’s about the fact that data sizes will keep increasing.

“We’re already automating the ability to provision machines in the cloud, and that’s the recipe for what we need. We have to take that same automation story to the Big Data governance story. We need Big Data governance tools that can automate Big Data governance, so we can deal with Big Data of increasing size.”

Those tools, however, aren’t necessarily out there right now, added Bloomberg. He said that 2014 should see a new effort from NoSQL database vendors toward supporting enterprise governance needs.

Indeed, governance of Big Data is a central topic of discussion in the mainstream media, thanks in part to Edward Snowden’s NSA revelations. “The more powerful the tools, the more likely someone will misuse them,” said Bloomberg. “We have to have an appropriate way for establishing the proper usage of those tools.”

Bloomberg was not the only one discussing governance and NoSQL usage in the enterprise. Andy Mendelsohn, senior vice president of database server technologies at Oracle, used his keynote to detail the benefits of Oracle’s NoSQL solution, which is based on the company’s Berkeley DB library for creating an in-memory key-value database.

Dave Segleau, director of product management for Oracle, has been working with Berkeley DB since he worked at Sleepycat, the company behind that database. When Sleepycat was acquired by Oracle in 2006, he joined the team that would go on to build Berkeley DB into an enterprise-grade NoSQL solution.

“You can think of it as the core technology we’ve been using for over 15 years for very fast, very efficient and simple data management, with a whole suite of manageability software around it,” said Segleau.

Mendelsohn announced that the Oracle NoSQL Database would be made available in a subscription model, and that it would be available for deployment into the Oracle Big Data Appliance.

“NoSQL has transitioned from being experimental, cutting-edge software, to being a fundamental piece of what enterprises are deploying into production,” said Segleau.
Transactional show
Another major theme at NoSQL Now was the rise of new transactional databases using the NoSQL name. FoundationDB, Starcounter and VoltDB were all demonstrated, each touting its own unique form of support for transactions.

Until now, most NoSQL solutions have specifically avoided supporting transactions, as they require much of the overhead and complexity most NoSQL solutions try to avoid. But FoundationDB was built from the ground up to support distributed transactions; Starcounter is specially designed for optimized transactions; and VoltDB goes as far as supporting full SQL within a NewSQL model.

FoundationDB was also at the conference to show off its newest acquisition: Akiban. Akiban was an independent NewSQL vendor, and this acquisition brings its SQL top layer onto the FoundationDB data layer.

Nick Lavezzo, cofounder of FoundationDB, said, “Now, instead of SQL support on FoundationDB being two years out, SQL support is a matter of months out. That’s going to further differentiate us from our competitors. We can offer key-value storage, and we’ll soon have document and SQL. Those will be many models accessing the same database, and able to do atomic operations that span all data and all data models in the cluster.”

VoltDB, on the other hand, has supported SQL from day one. This next-generation NewSQL database was co-written by Michael Stonebreaker, and other database luminaries.

Ryan Betts, field CTO of VoltDB, said that he’s expecting 2014 to be a year of consolidation within the NoSQL market. “I think there will be a flushing out. This year, we’ve started already. Next year, it will continue,” he said.

“The industry is going to come up with repeatable patterns. We have a LAMP stack; what’s the Big Data platform stack for this? We’re going to start seeing the selection of key technology that people will use to repeatedly build these stacks over and over again.”

Asa Holmstrom, CEO of Starcounter, is hoping that there is still plenty of room left for innovation in existing database-driven applications. Her company’s eponymous database can process extremely high numbers of transactions, yet relies on a single server model to simplify data management.

Holmstrom said that a beta of Starcounter 2.0 had been made available. In this new version, “Starcounter 2.0 is a little better version of the database, but it’s also a Web platform for client developers that have Web or application-based clients.”

New REST and JSON support are at the heart of this version, which Holmstrom hoped will expand the appeal of the database to more Web and mobile developers.