DataStax, distributor of the NoSQL database Apache Cassandra, yesterday announced the release of the latest version of its Enterprise distribution. As well as Cassandra, Enterprise bundles together a whole suite of Apache tools including the Solr search server and Hadoop big data platform, along with the DataStax OpsCenter administrative console, the latest version of which the company claims speeds up the display of monitoring statistics by tenfold.
DataStax CEO Billy Bosworth (pictured) told Computing that this version is all about making life easier for developers and architects.
"In this release we are focusing squarely on the developer community," he said. "We've completed the language interface CQL [the Cassandra Query Language], which is a subset of SQL, meaning that a traditional SQL developer will be very comfortable with the syntax and the language. We've also taken on some new drivers and made the operational side a lot easier as well."
This may sound like a familiar pitch to anyone who has followed the evolution of other NoSQL databases such as MongoDB and CouchDB. While each started from a different place, all now seem to be converging around a core feature set of tools for developer productivity, SQL-like querying, approximating ACID transactions, and latterly administration tools and improved security features. In other words, they are all replicating the functionality, ease of use and transactional integrity of their traditional SQL counterparts. What does Bosworth think will set them apart three years down the road?
"I can tell you what sets Cassandra apart. Our architecture was built from day-one for enterprise scale and modern online apps. We are differentiated from other players because we have a fully distributed system that is a master-less environment - every machine is the same type."
This architecture, he says, has consequences for the fault-tolerant, always-on, highly distributed systems DataStax specialises in, because a single database can be easily spread across on-premise and cloud data centres and across multiple geographical locations. Bosworth claims that Cassandra is the leader in this regard.
"What will keep us differentiated is the ability to solve those mission-critical, always-available problems at global scale. This is so fundamental to the architecture that to add it later would be like ripping and replacing the plumbing for an entire city," he said.
"One of our customers, OpenWave Messaging, lost a country. They lost Japan when the tsunami hit. They're a messaging backbone - in a critical event like that people turn to messaging, so they can't go just down when the local data centres stop working. They've had to create an architecture that spans most countries in the world, as many of our customers do."
Oracle and the innovator's dilemma
One thing that many of the the independent NoSQL players have in common is a desire to give Oracle a good kicking while it's, if not exactly down, at least looking warily over its shoulder. Bosworth says his company is taking business away from Oracle, and that the giant's core model is in trouble.
"[A couple of years ago] I noticed that Oracle had stopped reporting database revenues. When the world's largest database company stops reporting database revenues something's up.
"Oracle faces the innovator's dilemma: when you're a market leader, driven by your outlay on your cash-cow products and maintenance streams, what do you do when a disruptive technology comes along that's about a tenth of the cost?
"They are also married to their hardware play. I think it's understood that Sun was probably not the wisest acquisition, but they're in it and they need to work it out. For them their NoSQL database is in the context of their big data appliance, which means it's an Exadata conversation, you're talking millions of dollars. Our customers don't want to go down that path to 'Mainframe 2.0' - they want flexibility and freedom."
A narrowing gap
But is this really a revolution? Computing's own research demonstrates that while there is a lot of interest in big data, demonstrable real-world examples in the UK are still thin on the ground.
"When I used to work for an ISV, we used to say that for Europe there was an 18-24 month lag [behind Silicon Valley] in adopting new versions," said Bosworth.
"For NoSQL I'd say there's a 9-12 month gap. You have quite a few companies hitting the wall with Oracle, wanting to add new functionality or to scale up to new levels, and a smaller subset starting brand new projects. Away from Silicon Valley in the US we're seeing it spreading to mainstream bricks-and-mortar companies and this is happening in Europe too."
Putting its money where its mouth is, DataStax recently opened an EMEA office in Uxbridge, Middlesex, which currently has 15 staff and is hiring more. Bosworth said that part of a recently announced tranch of venture capital funding, totalling $45m, would be put towards further expansion.
"We'll be taking the funding and putting it towards that expansion as well as marketing and sales worldwide," he said.
Sometimes, the power of the mainframe is the most cost effective answer. Computing's Peter Gothard puts Computing's readers' questions on the future of the mainframe to IBM's Z13 expert Steven Dickens.
This Dummies white paper will help you better understand business process management (BPM)