Facebook alumni wunderkind MemSQL is making a lot of wealthy friends

Lucy Carey

Following an epic $35 million funding round, we examine the case for backing the company aiming to “bring the database to the masses”.

Big data continues to make headlines – both in terms of tech, and big payouts. Even with huge players like MongoDB lapping up the bulk of the attention, there’s more than enough room for smaller guys to get in on that market growth curve.

This month, MemSQL, the speedy wunderkind of ex-Facebook engineers Eric Frenkiel and Nikita Shamgunov, has recently enjoyed its own cash bonanza, having raised $35 million in new funding. Investors to date include the likes of Accel Partners, who led the latest funding round, as well as Khosla Ventures, Data Collective, and First Round Capital.

Officially launched in mid-2012, this in-memory database system claims to be “30x times faster than conventional databases on disk”. Back in their Facebook days, Frenkiel and Shamgunov had seen firsthand how hawk-eyed analysts scrutinised SQL queries to spot trends in users of the sprawling social network. MemSQL was born when the pair decided to develop a way to extend this capability to smaller companies without the budget for large engineering teams to tool on their databases.

By incorporating familiar elements such as SQL-style queries and the standard MySQL API, MemSQL can boost performance by storing data in main memory and processing it in parallel.

It’s ability to swiftly deal with both transactional and analytic workloads means that people like Comcast use it in front of Hadoop to add in all-important real-time capabilities. In the case of Zynga, it’s used in front of Vertica to add transaction support to the columnar system.

Whilst bigger, more established companies may have a death grip on the top of the market for now, there are countless startups like MemSQL looking to prise out a chunk of their own. And, along with it’s purported speed, there’s the small value of its budget friendliness.

MemSQL has a distinctly more friendly pricing structure than Oracle - Frenkiel notes that San Fransisco behemouth sets its prices at, “about $100,000 per terabyte (TB) when you break down its $3 million hardware cost”. Random Access Memory (RAM) on a commodity server priced at around $4,000 per TB.

He’s termed the technology “future proof,” given that, to increase capacity, all companies have to do is tag on a few additional cheap physical servers to their cluster, eschewing the need for pricier hardware add-ons.

So, faster and cheaper than the market leaders. What else? Well, there’s also the fact that there’s also no need for users to master a complicated new query language to get on board, making it an attractive option for anyone looking to switch systems with as little hassle as possible.

In an interview with Venturebeat, Frenkiel explained that this latest cash injection will go towards driving further sales, and evolving the database. To date, the startup has raised a total of $45.2 million, and attracted users such as Comcast, Shutterstock, Ziff Davis, and Zynga. And did we mention that Ashton Kutcher thinks they’re pretty sweet, dude?

By 2017, the big data technology market is expected to grow by $32.4 billion – that’s  about six times faster than the growth rate of the information and communication technology market as a whole. Even with the shadow of Oracle looming over them, that’s more than enough room for MemSQL to continue making rain from both teeny little startups in the shallows, as well as some very big fish indeed.

Inline Feedbacks
View all comments