DataStax, a San Mateo, Calif.-based startup offering a commercial version of the Apache Cassandra NoSQL database, has raised a $45 million series D round led by new investor Scale Ventures. Existing investors Lightspeed Venture Partners, Crosslink Capital and Meritech Capital Partners, and new investors DFJ Growth and Next World Capital, also contributed. DataStax has now raised a total of $83.7 million in venture capital since launching in 2010 as Riptano.
You might not know it from browsing the headlines or examining the software stacks of up-and-coming web applications, but Cassandra still exists and is actually quite popular among a certain class of user. Sure, the database has had its share of bad press over the years, MongoDB still gets all the attention, and many of the cool developers — including those at Cassandra birthplace Facebook — have chosen HBase as their BigTable clone du jour (here’s a nice overview of the NoSQL landscape for the uninitiated). But Cassandra has been steadily improving thanks in part to DataStax’s efforts to lead its development.
Even if Facebook and Twitter (this time for real, I’ve heard) have stopped building applications on Cassandra, we’ve covered its continued use in companies such as eBay, Netflix, Eventbrite and BloomReach. There’s also a London-based company called Acunu, whose Founder and CEO Tim Moreton will be presenting at our Structure: Europe conference in September, that has built a real-time analytics platform on top of Cassandra. For its commercial version of the database, called DataStax enterprise, DataStax claims more than 300 customers, including 20 companies in the Fortune 100 (one of which is Disney).
Making hay one Oracle customer at a time
Cassandra’s success with such large users has to do with its ability to handle large-scale online applications that demand steady levels of performance, DataStax CEO Billy Bosworth told me. Scalability and performance have never been among Cassandra’s shortcomings, and the database is capable of replicating data across data centers. Large companies used to choose Oracle for applications that needed these capabilities, but now that NoSQL options are around and relatively mature, companies are rethinking whether the relational database model was ever really correct for some applications in the first place.
And a good number of them, Bosworth said, are deciding it wasn’t and switching to Cassandra. When he joined DataStax in 2011, he thought NoSQL adoption would mirror that of client-server adoption, meaning a handful of customers moving to Cassandra from relational database by 2014 would be a sign of success. Everything actually happened much faster: “The scale I had anticipated for 2014 actually hit us in 2012,” Bosworth said.
In fact, he added, the company wasn’t even out raising money; investors came to DataStax with a lot of money at a good valuation, so it jumped on the opportunity. The new capital will help the company expand its European presence as well as generally ramp up product development and marketing.
Open source FTW!
He credits the open source nature of many NoSQL databases with helping speed the pace of adoption beyond what previous technological shifts have experienced. That makes sense. The monetary barriers are so much lower when it’s free (at least in terms of licenses) to get started and scale something that developer time and curiosity might be the biggest obstacles to trying new things that aren’t being offered as part of a commercial software product.
Of course, DataStax hasn’t exactly hurt its own cause. The company maintains a free Community Edition of its Cassandra distribution that also includes a lightweight version of the company’s management software, and then it has the full-featured Enterprise package that’s prepackaged with Hadoop and Solr for search. Both of DataStax’s products received substantial upgrades as of Tuesday, coinciding with the release of Apache Cassandra 2.0.
But even though DataStax has raised a lot of money and landed some big-name users, Bosworth isn’t so blind as to declare victory over the database market. He thinks we’re in an era of “polyglot persistence” (which is the database version of polyglot programming) in which developers choose the right database for the right job. Based on how many companies are running Cassandra, a relational database, and even other NoSQL stores such as MongoDB or maybe HBase, he’s probably right.