DataStax Rakes $45 Million; Schemes Growth
While Hadoop has been getting the elephant’s share of attention recently in NoSQL database circles, Cassandra database vendor, DataStax, has been dutifully squirrelling away at their own plans, which it was revealed this morning will be fueled by a $45 million dollar series D funding round.
The new funding round is being led by a VC firm new to DataStax, Scale Venture Partners, and includes participation from other new investors, DFJ Growth and Next World Capital. Existing investors, including Lightspeed Venture Partners, Crosslink Capital and Meritech Capital Partners also kicked in.
The new round tags on to the $39 million that DataStax has managed to collect to date since it first launched as Riptano in 2010. With the new haul, DataStax has raked in roughly $84 million in total investment dollars in that time, with the overwhelming bulk of that ($81 million) being marshaled during the reign of CEO, Billy Bosworth.
In addition to the funding round, DataStax today also announced the release of the latest version (3.1) of their keystone database software, DataStax Enterprise (DSE), based on Apache Cassandra. The release contains upgrades that widen the data per node (up to 10x more than previous versions), and introduces a faux-SQL query language, Cassandra Query Language (CQL), aimed at widening the on-ramp for the DataStax database software.
We spoke with Bosworth about DataStax plans for growth recently as the company was preparing for their annual Cassandra Summit. Bosworth was eager to tell us of their claimed success that DataStax was having in their quest to supplant Oracle as the database of choice in the enterprise.
Bosworth has good reasons to crow. Since May 2011, when Bosworth first joined the company, he says they’ve grown their customer base from 27 to over 300 at present. “What’s more exciting than the raw numbers is the types of customers,” he boasted. “We are already in 20 of the Fortune 100, which is faster than I thought, to be honest with you.” In that time, the company has grown from 26 employees to over 100, which Bosworth says they plan on doubling in the next 3 to 4 quarters. “Hiring is foot to the floor – we’re hiring as quickly as we can in multiple locations.”
Included in those locations is a new full subsidiary in London that was opened earlier this spring to address what Bosworth called “demand we are seeing from the international community.”
The new funding haul gives a significant amount of credibility to the company’s growth claims as well as its overall standing in the advancement of NoSQL database technologies into the enterprise. Bosworth says that the success that they’ve had with companies like Netflix, Ooyala, and Openwave Systems has turned into wildfire in terms of word-of-mouth referencing as these companies speak openly about their transitions away from Oracle and towards DataStax and Cassandra.
“When Netflix tried to make Oracle scale to the geographical expansion that they needed and the performance that they needed, they just got tired of fighting that battle, and they said, we’ve got to find a way to start a migration off of Oracle,” explained Bosworth, adding that the company is about 95% completely there with only back-end office stuff left in the relational database.
Bosworth pointed at the growth and growing depth of the annual Cassandra Summit, which took place last month, as public-facing proof of the growth (and growth potential) of the Cassandra-sphere. Participants at the event (both speakers and sponsors) included an array of familiar names, including Sony, Walmart, Intuit, eBay, Microsoft, Splunk, Fusion-io, Accenture, Instagram, Dell, Spotify, Splunk, and others.
The popular knock in recent times against Cassandra vs. it’s NoSQL competitors has been that Cassandra is relatively more difficult to use – which Bosworth says the company spent a lot of resources in combatting through documentation that was not available when they first got underway with companies like Netflix. DataStax CTO and co-founder, Jonathan Ellis tells us that the work they’re doing now puts some focus on turning that perception around.
“If you kind of oversimplify the NoSQL landscape, you have MongoDB is easy to use, Cassandra scales, and HBase is for Hadoop users…so we have a reputation for performance, scalability, and reliability, but not so much for ease of use,” illustrated Ellis.
In their bid to address this, DataStax released their new CQL query language today as part of their 3.1 announcement, giving developers the opportunity to recycle their relational database knowledge for use in the NoSQL world of Cassandra. The move, says Ellis, is a bid to get developers who have not been as eager “to live on the bleeding edge” a foundation of comfort and familiarity. “The early adopters came from scaling and reliability problem environments, so they were willing to put up with a less user friendly developer environment,” he said, adding that with moves like the addition of CQL (and the addition of copious documentation), DataStax is trying to “cross the chasm” into being a more mass market play.
In addition to an ease-of-use upgrade, the new DSE version includes an optimization which the company says will allow up to 10X more Cassandra data per node than previous versions. Previously, says Ellis, they were limited to about half a terabyte to 1 terabyte per machine, but with the 3.1 release, they’re looking at data ranges of five to ten terabytes per machine.
“The limitation there is that Cassandra keeps a lot of metadata about what you stored in it in the Java heap, and because of the limitations of the JVM garbage collection, you can’t push the heap past eight to twelve gigabytes – you really start running into problems once you get past that size. So what we’ve done for DSE 3.1 is move the structures. We’ve moved the metadata off heap, and manage it directly in native memory so that we’re not limited by the JVM garbage collector throughput.”
With these optimizations in place and their cash arsenal restocked, Bosworth says that they’re not standing still. “If we can continue to make customers successful the way that we are, and the type of customers that we’re making successful, then I have no doubt in my mind that we will continue to keep innovating and meeting their needs.”
“A big part of that is staying focused on what you know and who you are. We don’t try and do 50 different things – we try and do our thing extremely well because that’s an environment that isn’t going away. The enterprise online application database is going to be a staple in all enterprises – just like it’s been for the last 20 years, it’s probably going to be for the next 20.”