February 19, 2014

How Actian Plans to Take Over the Big Data World

Alex Woodie

One company that has been moving in the big data analytics space is Actian. Formerly known as Ingres, the Redwood City, California company has spent $200 million over the past few years assembling an all-star cast of analytic technologies, including a pair of fast analytic databases, VectorWise and ParAccel, and building the connective tooling that let them work in the real world. Now Actian CTO Mike Hoskins says all the pieces are in place, and the company is looking to make a big splash in 2014.

Founded in 1980 as Relational Technology Inc., RTI competed against the likes of IBM, Oracle, Microsoft, Informix, Sybase, and Postgres in the relational database world. In 1989, the company took the Ingres name. It got caught up in Computer Associates’ acquisition net in 1994, and was spun out as an independent company in 2005, with backing from private equity. 

Things started to heat up in 2010, when the company bought VectorWise, an analytic SQL relational database that was developed at Centrum Wiskunde & Informatica (CWI), the Dutch National Research Institute for Mathematics and Computer Science. As the big data market was heating up, so did Ingres, which changed its name to Actian in September 2011. The newly named company added a NoSQL object database to the mix in 2012 when it acquired Versant, which had rejected an offer by Unicom Systems.

In 2013, Actian bought Pervasive Software, which owned a SQL relational database and an array of ETL and data integration tools. Also in 2013, Actian nabbed ParAccel, which had developed a very fast and scalable columnar analytic database that had smoked the competition on benchmark tests so thoroughly that Amazon picked it to be the technology behind its Red Shift data warehousing as a service offering.

“We’re in a highly transitional period,” Hoskins says. “These are intra-generational shifts that come along only once every 30 years. We know that we won’t be using in the next 30 years what we used in the last 30. It’s kind of a Wild West period right now. There’s new vendors emerging left and right, so we put our stake in the ground.  We made a $200 million investment… in these various advanced next generation stacks.”

That stake involved buying or building much of the big analytics tooling that customers will need as they obtain their “massively parallelized, super-scaling modern software stacks,” Hoskins says. “Not stacks like the 10, 20, or 30 year-old stone age software stacks that most customers are using today, which are utterly underpowered and will be overwhelmed in the age of data as big data scales in complexity.”

Hadoop is part of Actian’s vision, as are the various NoSQL databases that are cropping up to server operational and analytics needs. Actian’s strategy is to surround its customer’s Hadoop data lake with tools that enrich the software and make it more useful.

“Hadoop is beautiful infrastructure. It’s very young and very immature. It lacks a tremendous amounts of efficiencies and optimizations and price performance,” Hoskins says. “I discovered it six years ago doing heavy research to make sure that all our solutions co-exists and exploit the goodness of Hadoop. But we see it as a layer of infrastructure that sits below us.”

While Actian invested a lot of time and money in developing a complete analytics stack, it’s not locking users in or out. Some Actian customers use its ETL product DataFlow with other analytic database, while others use third-party ETL tools with its analytic databases, Vector or Matrix.

“It’s a platform in the sense that there are best-of-breed technologies that we spent $200 million to acquire and we’ve now arranged them together as the natural starting place,” Hoskins says. “But it’s not the ending place. It’s not just one size fits all….It’s very important to interface with best-of-breed [products]. If you look at some of the legacy vendors, the IBMs, Oracles, HPs–they have great difficulty running fast enough to love open source and interface to that best of breed like we’ve done.”

Graph processing and semantic Triple Stores are of particular interest to Hoskins at the moment. Actian is the primary investor in SPARQL City, a startup founded by Netezza co-founder Barry Zane (former ParAccel CTO) that’s developing a semantic tripe-store that Actian will interface with.

Actian CTO Mike Hoskins

“These are fascinating things,” Hoskins says. “It’s a very rich, young, emerging ecosystem.  We’re big believers in NoSQL. Just like the Hadoop question, the NoSQL change is extremely good for the industry. But it needs visual tooling. Just like the early days of SQL database had no loaders, no unloaders, low performance, no sort, no optimizers, no visual tooling. No nothing. It’s incumbent on Actian to take things like NoSQL and Hadoop, industrialize them, and build the visual tooling.”

The company has done the development work to stitch together the four “pillars” of its product stack, and is now ready to take this software to market. “We chose 2014 to make a loud statement,” Hoskins says.

Actian had $140 million in revenues last year, and will go well past that this year. The company has more than 10,000 customers, including some big companies like Home Depot, OfficeMax, GE Power and Water, Sheetz, Intuit, and Autometrics.

Related Items:

Big Data Market Growth Exceeds Expectations 

Amazon Goes SSD with RedShift ‘Dense Compute Nodes’ 

Actian Teams with Hortonworks 

former ParAccel CTO)

Tags: , , ,