MemSQL Giving Away Database for Four-Node Clusters

(Marynchenko Oleksandr/Shutterstock)

Everybody likes free stuff. Today MemSQL announced that it’s getting into the free software market by giving away copies of its distributed relational database management system for clusters up to four nodes and 128GB of RAM.

The freebie database comes with all the security and resiliency features that enterprise customers pay for, and is limited in no other ways than the number of nodes and total memory used, according to MemSQL co-founder and CEO Nikita Shamgunov.

“What can you do on four nodes? Well, you can actually do a lot more than let’s say what you get from a legacy software vendor like Oracle or Microsoft,” Shamgunov tells Datanami. “Customers to date are doing major production work on sub four-node, sub-128GB RAM clusters.”

MemSQL positions its database as a next-generation distributed relational database – what the market termed “NewSQL” for lack of a better word –designed primarily for demanding SQL workloads. However, it also brings some analytical capabilities, courtesy of a built-in column store, that enables it to power real-time dashboards. While the database ostensibly is “in-memory,” it can also store data on spinning disk and SSDs, with a total capacity of about 1 petabyte, Shamgunov says.

The San Francisco company has landed some blue chip accounts since launching its distributed database five-and-a-half years ago, including Kellogg’s, Samsung, and Comcast. While MemSQL continues to chase big enterprise users that bring in $1 million or more in license fees every year, the company is also cognizant of the need to seed its growth, hence today’s introduction of a free tier with version 6.7.

Shamgunov realizes he’s giving up some license revenue today with the free database offer, but he’s confident MemSQL will grow faster in the long term as a result.

“We certainly will sacrifice a certain amount of revenue, but our customers are giving a signal to where the real money is, and that’s above the four nodes,” Shamgunov says. “If you hook them on the simplicity and elasticity and scalability of the software, they certainly see value, and four nodes is enough to understand that very well. They say, ‘OK, my bigger workloads will require eight nodes and then 16 nodes and 20 nodes, or in the case of Uber, 200 nodes.'”

MemSQL is privately held and is in venture-fueled growth mode. The company, which attracted $30 million in capital earlier this year, says its year-over-year bookings grew by 200%.  That growth was partly driven by a 3x increase in add-on revenue, or sales that occur after a company has stated using the database.

The company is introducing other features with MemSQL version 6.7, which debuts today, including MemSQL Studio, a new monitoring tool designed to make it easier for developers and DevOps professionals to understand what’s going on with their cluster; and up to a 100x performance boost for star join queries.

Related Items:

Pandora’s Journey from Hadoop to MemSQL

MemSQL Gains Traction, New Investors