Last year, we introduced you to MemSQL, a plucky young Y Combinator alum that was building technology that would let developers give their databases a Nitrous boost, while simplifying application development and maintenance. Founded by ex-Facebookers Eric Frenkiel and Nikita Shamgunov, MemSQL raised $2.1 million in seed funding last July from an impressive roster of investors, before going silent to plug away at their private beta.
After a year of development, MemSQL is officially ready for primetime. Today, the startup launches its next-gen database into the world, and it's got some shiny new coin to go with its new product, bringing total investment to $5 million. As part of the follow-on seed funding, existing investors First Round Capital, SV Angel, Y Combinator, Paul Bucheit and Ashton Kutcher are joined by IA Ventures, Max Levchin, Aaron Levie and Data Collective - to name a few.
The startup intends to use its new capital to scale its infrastructure and take advantage of areas of the market where fast analysis of machine data is crucial, like financial services, digital advertising technology, and telecom and mobile services.
Those familiar with the space may wonder what makes MemSQL so special, out of the multiple VC-backed, already-been-around-for-awhile “newSQL” startups out there doin' their thing. (Like Clustrix, GenieDB, VoltDB, RethinkDB, ScaleDB, JustOneDB, Tokutek, Akiban, CodeFutures, ScaleBase, etc.)
Well, because both Frenkiel and Shamgunov were previously database guys at Facebook, and they've taken some parts of FB's approach and repurposed it for speed and flexibility - and a wider audience. It also helps that Shamgunov spent 6 years as a database engineer at Microsoft working on SQL, has several patents to his name, and is a world medalist in computing machinery contests. We're still not really sure what that means, but it sounds impressive.
The co-founder's Microsoft experience also helps MemSQL, because essentially what these guys are trying to do is offer a cheaper, works-for-everyone alternative to what the big players like Oracle and MSFT offer for the enterprise at a much higher price point.
But before going any further, for those unfamiliar, the “newSQL” movement is basically trying to build a better replacement for legacy SQL databases, like MySQL, which tend not to play so well with all these modern, newfangled web applications. They struggle to keep up with these data producers, when that, after all, is their job.
What's more, big data is big and will only get bigger, and developers and companies need better ways to manage that data, control it, and get the most out of it. For companies that operate businesses on the Web, it's imperative that when they get hit with a lot of traffic or once they move all their data etc. into the cloud (which can make for slower through-put thanks to bottlenecks) that they be able to fall back on memory storage to keep their apps or operations up and running. Big companies like Facebook and Twitter have the engineering resources to be able to do this efficiently, but most companies don't.
MemSQL offers a database that is relational and distributed, and tweaks the way data is put into memory, with the right structures and access patterns, so that apps can functionally use SQL at scale without having to use Memcached or NoSQL.
For those who just fell asleep, basically that means less engineering overhead - less work for developers. This is ideal, Frenkiel says, for apps that need to process machine data at hyper speed, as MemSQL claims to be able to increase an app's through-put (and ability to retrieve actionable insight) by 30x while still offering SQL and flexibility.
“Quickly sifting through actionable data has become a competitive advantage in the enterprise, particularly as organizations are saddled with increasingly large and complex data sets,” Frenkiel says. “MemSQL was designed to tackle Big Data problems by accelerating an application's throughput while still offering SQL, unlike other solutions which tend to either be too slow or too limited in functionality."
MemSQL at home here and product demo below: