MemSQL Readies Cloud Database Service and Revved Operational Analytics Engine

(Pictured above: MemSQL’s Peter Guagenti introduces Helios at the Strata Data Conference.)

MemSQL is expanding its operational data analytics platform to the public cloud with the introduction of an on-demand service and the forthcoming release of its revved up on-premises distributed database engine.

The company unveiled MemSQL Helios, an on-demand database-as-a-service at last week’s Strata Data Conference in New York, and said the second beta of the MemSQL 7.0 core database is now available for download. The new MemSQL Helios database as a service is now in private preview in AWS and Google Cloud Platform. MemSQL plans to offer it in Microsoft Azure next year as well. The new cloud version eliminates the need for script deployments or VM management but operates the same as the flagship MemSQL database system.

It also allows hybrid implementations using the company’s recently introduced MemSQL Kubernetes Operator, or directly within the Helios service. Peter Guagenti, MemSQL’s CMO said the move made sense now that nearly 70% of MemSQL implementations run as VMs or virtual servers in the public cloud. The release of the Kubernetes Operator was a clear signal that MemSQL planned to offer a cloud-native version as a service, which analysts said is where the market for future applications is headed.

“MemSQL Helios offers the full capabilities and features of MemSQL with no changes but deployed instantly and elastically,” Guagenti said.

Because it supports Kubernetes orchestration, MemSQL Helios will run in multicloud environments, Guagenti added.

“We’re leveraging a lot of the innovations of Kubernetes, which makes it easy to deploy MemSQL as a service across cloud providers,” he added. “Just click a button to deploy a cluster, pay on consumption and go from there.”

dbInsight’s Tony Baer

Industry analyst Tony Baer of dbInsight said the move to the cloud makes sense. “It’s what every database player has to do and is, or is in the process of doing,” Baer said. “Their mixing of OLTP and analytics, pairing row plus column stores — well they are by far not the first ones out of the chute with that. In fact, SAP pulled this off years ago with HANA, where instead of having physically separate stores, it makes them virtual.”

But Guagenti said the company’s core relational database engine offers more than 10 times the performance of an Oracle Exadata or SAP HANA databases at a third of the cost. Guagenti said half of the world’s largest banks use MemSQL including a list of noteworthy customers such as Akamai, Comcast, Sony and Uber.

Nevertheless, MemSQL is a relatively small player, with about 200 employees and approximately 130 customers. Last year, MemSQL announced its Series D funding round from GV (formerly Google Ventures) and Glynn Capital. With funding from existing investors, including Accel, Caffeinated Capital, Data Collective and IA Ventures MemSQL has raised $110 million.

In addition to bringing in Giagentl, MemSQL brought Microsoft database engineering veteran Ted Kummert to its board and hired Cloudera and Hortonworks veteran Jason Wakeam to expand MemSQL’s partner base. Wakeam told Channel Futures that…

MemSQL’s Jason Wakeam

…MemSQL has 50 partners worldwide.

“A good half of those are productive with us in pipeline today, either actively registering deals or closing opportunities in that pipeline,” Wakeam said. Additionally, MemSQL has partners that provide service delivery, he said.

Wakeam added that MemSQL has launched a range of recruitment activities, both domestically as well as internationally, with various outreach activities. “The partners that we tend to find the most productive are those that are working with the major cloud vendors to provide both operational data management and analytics as well as cloud migration services to those clients,” he said. The latest releases will give partners a stronger story and solution development opportunity, he added.

The new MemSQL 7.0, set for release later this year, offers two key new features. One, called SingleStore, has a new data management storage engine that will process operational and analytical workloads using a single table type, which the company said will deliver major scalability improvements. The other feature is faster replication speeds for Tier 1 workloads and providing systems of record.

Pund-IT’s Charles King

“MemSQL V7 combines some significant benefits, including the company’s long time focus on [transactional and analytical] and in-memory functions that simultaneously support transaction and analytics processes,” said Pund-IT principal analyst Charles King. “The new Single Store capability eliminates the need to choose between a row or a column store for different workloads and offers incremental backup and synchronous replication functions that impose little if any performance penalty.”

While King said these are innovative capabilities, it remains to be seen whether it’s enough to expand the company’s footprint. “That isn’t because the company is unworthy in any way,” he said. “It simply reflects just how sticky enterprise databases tend to be.”