Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
On the surface, Eric Frenkiel hasn’t accomplished a whole lot, at least by Silicon Valley standards. He graduated from Stanford in 2008 with an industrial engineering degree, spent a little time at Facebook as a sales engineer and then left to start his own company — run-of-the-mill stuff. Another way to look at it, though, is that Frenkiel is co-founder and CEO of MemSQL, one of the hottest database startups around.
However, when Frenkiel and co-founder and CTO Nikita Shamgunov decided to leave Facebook in 2011, with a near-term IPO a foregone conclusion, people thought they were crazy. It was probably a fair assessment. After all, their plan was to get into Y Combinator, a startup incubator that wasn’t and still isn’t known for its enterprise startups, and that paid a mere $17,000 a year.
Making matters worse, Frenkiel says, “We applied with zero code, just an idea.”
Frenkiel credits Shamgunov’s technical chops with helping them earn enough trust to get accepted. It’s easy to see why: Shamgunov, a computer science Ph.D., helped scale out Facebook’s MySQL cluster and spent six years as a distinguished senior database engineer at Microsoft. Frenkiel is probably selling himself short, though.
He’s on this list because he’s the quintessential Silicon Valley startup CEO. He has technical knowledge, big ideas and when you hear him talk about his company’s technology, you want to believe. He says he’s always been just as comfortable speaking with salespeople as with engineers, and that he avoided a traditional computer science education in part because he’s a naturally social person.
(Listen to him talk about the staying power of SQL and you get the picture. “It’s about 150 words,” he waxed during our call, “but, man, what complex expressions you can create.”)
Taking cues from Facebook
Regardless how the credit is split, though, Frenkiel and Shamgunov also came to the table with a can’t-miss idea. They wanted to turn what they had already seen at Facebook — massive databases running in memory powering everything from user transactions to analytics — into something that’s repeatable within more-traditional data centers.
“I think real-time analytics is the next frontier in big data,” Frenkiel said. “It’s our belief that any successful company will have to become a data-driven company to survive.”
When Frenkiel and Shamgunov asked themselves how they would build a database from scratch to solve those issues, they came up with an answer. “We wanted to rethink basic assumptions about hardware,” Frenkiel said. “It’s just a whole new world in terms of what people can actually do with their hardware.”
So they built a database designed to take advantage of cheap, plentiful memory and commodity servers housing dozens of 64-bit CPU cores. They even took a page from the Facebook playbook (its HipHop tool, to be exact), by transforming SQL into C++ so machines can read it faster.
Success comes to those who take risks
Today — just a year after launching publicly — MemSQL is a full-on distributed database with monitoring tools, online replication, high availability and the whole nine yards. Zynga is using it in production to do real-time analytics on its social-gaming platform. Morgan Stanley has 25,000 financial advisers using a MemSQL-powered application for real-time bond analysis.
Frenkiel is already thinking about the future. He’d like to help them out, perhaps by building a sort of SQL-meets-Hadoop system that takes advantage of advances in memory to provide different levels of speed and reliability. He’s excited about the advent of 64-bit ARM processors that could revolutionize distributed systems by letting companies put 5,000 low-power cores on a single rack.
He and Shamgunov might have missed out on cashing in on Facebook’s IPO, but they’ve done well for themselves. Heck, a “very, very, very early version of MemSQL” they built to help Y Combinator classmate LikeALittle scale from zero users to 20 million in six weeks so impressed LikeALittle investor Ashton Kutcher that he contributed to the $2.1 million in seed funding MemSQL raised just three months after forming.
“It turns out we were only crazy by half,” Frenkiel joked. (Although, he admits, “I thought it was a prank call [when Kutcher rang]. I was getting ready to hang up.”)
Beside, he tries to live by the motto that you never want to live to regret the things you didn’t do. “If I didn’t start the company,” he said, “somebody else would have done it.”