For a 12-person startup, San Mateo-based Groovy Corp. is causing a fair bit of controversy in Silicon Valley. The reason? Groovy claims to have developed the capacity to do what has eluded all others — truly real-time data relay through its proprietary in-memory relational database management system, called the SQL Switch.
Truly real-time? Aren’t startups like Twitter doing that already? Actually, no, says Groovy’s CEO, Joe Ward. “If you go to Ticketmaster, do you see the tickets disappearing as they are sold? You don’t know if you’ve won or lost an auction on eBay until you’ve refreshed your page. Even on Twitter, you are seconds behind real time,” he says, calling it a “computer science problem that no one has really been able to solve.”
At least, no one has solved the problem cheaply or efficiently. Major media companies like Bloomberg and peHUB parent Thomson Reuters deliver some of their data in real time, but Ward says that their deep pockets allow them to “bolt together different, costly solutions to drive customers to a real-time-like experience.” The rest of the more than 99 percent of other data on the Web experiences a lag of anywhere from minutes to seconds, as browsers systematically retrieve updates from the servers with which they are communicating.
Enter Groovy, a three-year-old company cofounded by Ward and Ray Huetter, an Australian PhD who Ward says developed Groovy’s core technology more than 10 years ago — before the world began going bananas over the potential of real-time data transfers online.
The company’s biggest selling point is an extension to the database computer language SQL that allows Groovy’s platform to push data out to users, rather than force their browsers to repeatedly poll — and stress –corporate servers, particularly those dealing with high-volume traffic.
“Polling is like saying, ‘Are we there? Are we there yet?’” says Ward. “We take the opposite approach; we proactively say, “Here, this is where we are right now. You needn’t keep asking.’”
Groovy’s performance claims are bold, to say the least. Ward says that its special Intel boxes can cut costs for transactional data by 37 times (by reducing server load) and that they can reduce the response time of clients by 105 times. (The company claims that each box can handle 200,000 SQL operations per second.)
Ward also tells me that Groovy’s performance is constantly being tested by the R&D department of Intel, which is helping it benchmark its technology against traditional databases for a 6 percent stake in the company.
The company has at least one believer: in December, MySpace announced that Groovy was one of two launch partners when it announced a new suite of APIs that invite third party developers to play around with the data that gets published on its network. According to a trusted source, Groovy is also working with a major digital media company, though Ward declined to comment on that relationship.
Yet not everyone is buying what Groovy is selling, including companies that would seemingly benefit the most from its advancements. Ward says that angel Ron Conway introduced his team to Twitter 18 months ago, when Groovy was just beginning to package its technology. “They thought we were early,” says Ward, who adds that when they talked again with Twitter six months ago, they were told, “We wish we’d met you two years ago.”
Twitter has “such a high-profile now that every vendor is telling them, ‘We can do this, we can do that,’” says Ward. “I’m not concerned because we’re working on something exactly in that space that’s very good, and when it gets announced, we won’t have this [credibility] problem.”
To date, Groovy is bootsrapped; Ward says that a Series B round with a corporate VC other than Intel is in the works.