MemSQL Launches Next-Generation Database

A screenshot of MemSQL processing a workload.

Silicon Valley is filled to bursting with entrepreneurial types who left an established company to start their own little firm. MemSQL is no different, having been founded in 2011 by a pair of former Facebook engineers, Eric Frenkiel and Nikita Shamgunov, who sought a new challenge.

“Sometimes it’s scary to leave a big corporation,” Shamgunov, wrote in a June 18 posting on the MemSQL Developer Blog, “but the truth is that apart from fun, market pay, and the potential of a huge upside, a good early stage startup gives the kind of experience that makes a Top Coder extremely relevant in today’s tech industry.”

That’s part of MemSQL’s pitch to draw in qualified engineers to work on algorithmic and systems-level problems, particularly those related to distributed systems and cloud infrastructure. But what does MemSQL actually do?

As of June 18, the company offers a new eponymous database designed to accelerate the crunching of massive amounts of data, by placing a relational interface within an in-memory data tier. A video posted on MemSQL’s homepage offers a head-to-head comparison between MemSQL and MySQL, with both tackling the same workload of thousands of queries per second.

As demonstrated onscreen, MySQL pushes 3,000 queries per second, with performance degrading over time; MemSQL pushes 80,000 queries per second on the same machine while maintaining performance. For each unique type of SQL query, MemSQL (which is fully ACID-compliant and MySQL compatible) generates and compiles C++ code that runs against memory-optimized, lock-free data structures.

In essence, MemSQL claims the software delivers all the “benefits of SQL at NoSQL speeds,” including faster retrieval and analysis on massive datasets, a “flat learning curve” for MySQL users, and real-time analytics on data in flux.

Speed is good for businesses, along with the ability to crunch a rising tide of data. “Organizations are increasingly looking for IT solutions that can address the challenges associated with the kind of extreme information management Big Data requires,” Merv Adrian, a research vice president at Gartner, wrote in a June 18 statement tied to the release. The increasing need for business analytics, he added, will drive significant investments in information infrastructure over the next three years or so.

If every other week sees the release of another database product like MemSQL, then, it’s because of that growing market for mining the tons of data flooding small businesses and enterprises all over the world. Despite its sizable investment and the combined pedigree of its executives, the challenge facing the company will be to stand out in an environment crowded with other SQL-related startups.

Image: MemSQL