Database technology, while never the sexiest area, is attracting more money as legions of startups bet they can address a new class of big data-driven challenges—and dethrone database giants like Oracle, IBM, and Microsoft.
San Francisco-based Splice Machine topped off an $18 million Series B round earlier this month to advance its novel database, which it hopes corporations will turn to as a replacement for relational databases from incumbent providers. It’s just the latest in what’s been a lucrative year for entrepreneurs with clever ideas on how to architect the database of the future.
Investment in two fast-growing database technologies—Hadoop and NoSQL—has spiked over the past year and a half, attracting giant-size rounds and top-tier investors, according to data from CB Insights. Hadoop software uses a distributed file system that allows computing jobs to be run across many servers and NoSQL is a scalable alternative to relational databases. Venture funding for Hadoop companies grew 50 percent last year in dollars and 30 percent in deal volume, while three deals this year brought in a total of $1.1 billion, including a whopping $900 million round for Cloudera.
Investors and entrepreneurs say that the flurry of activity in databases—a field that seemed commoditized and dull in the mid-2000s—stems from the demands of big data applications. In the past, it may have not been worth saving and analyzing certain sources of information, such as sensor data or log data from IT equipment. Now, there’s a race to build databases and software systems that collect all that data and make sense of it using cheap server hardware.
Splice Machine is among those hoping to displace Oracle and others with a completely new database design for emerging computing problems. Its database allows programmers to use a common method for querying databases called the structured query language, or SQL, a core feature of relational databases. Splice Machine lets people run standard SQL-compliant applications and tools on a Hadoop cluster of servers.
The company’s business model is to sell its database software to customers who need a bigger database but want to run it on cheap commodity servers, rather than big, pricey servers from the likes of Oracle and HP.
“We really exist because of this fundamental shift that is happening with IT,” says Splice Machine CEO and co-founder Monte Zweben, a serial entrepreneur and former NASA artificial intelligence researcher. “The amount of data and the velocity of data coming from a variety of applications is outstripping the current typical technology to manage it—the relational database management system.”
There are a number of other companies trying to do similar things, because SQL compatibility makes database technologies more welcoming to corporate IT departments. Among them are HortonWorks, Cloudera, and MapR, notes Chris Lynch, a partner at Atlas Venture, who funded Hadapt, an analytics startup that was bought by TeraData last month.
In theory, integrating the SQL language closely with Hadoop makes a lot of sense, but technically it’s very hard to do, particularly for very distributed systems, says Mike Driscoll, CEO of MetaMarkets and an investor at the Data Collective, a Bay Area-based investment fund dedicated to big data.
Still, so many companies are trying to reinvent the database in various ways because billions of dollars are spent every year on database software, which is very difficult to displace once installed. It’s an increasingly crowded field, which means a new giant company may emerge from the pack but many will certainly fail trying.
“We need new architectures, which is why there’s this burst of innovation happening now,” Driscoll says. “And database companies are valuable because data is so central to everything an organization does. It’s the institutional memory, so it has its own gravitational force.”
