The Big Data market at large may not have heard of Actian Corporation, but in time that will likely change. Actian began life as Ingres Corporation, Inc., a spinoff from CA, formed as the latter divested itself of the Ingres relational database, an important early player in the RDBMS category. That may sound a bit ho-hum, but the plot thickens nicely.
In 2010, Actian acquired Vectorwise, formed around the intriguing vector database of the same name that commercialized much of the technology in the MonetDB database project. Then, in April of this year, Actian acquired Pervasive Software, makers of the DataRush data integration platform and RushAnalytics data mining suite. Then, just 2 weeks later, Actian acquired ParAccel, the Massively Parallel Processing (MPP) data warehouse that can run on commodity hardware, and whose technology powers the Amazon Redshift service.
DataRush creates data integration workflows that run in a parallelized fashion on a single server. Vector databases, including Vectorwise, optimize usage of modern CPU technology, including single instruction, multiple data (SIMD) techniques, to parallelize data processing within a single machine. MPP data warehouses like ParAccel, meanwhile, parallelize amongst multiple machines in a cluster.
These are very valueable pieces to have under one roof, as they together optimize operations between and within compute nodes, rather than obsessing on one model or the other exclusively. The question is how these pieces fit together.
From data integration to product integration
The answers to that question start to emerge today as Actian announces the Actian Cloud Platform and the ParAccel Big Data Analytics Platform.
The Actian Cloud Platform is the successor to Pervasive's Data Cloud, which runs on Amazon Web Services' EC2 infrastructure and is focused on handling data integration workloads elastically and to scale. The ParAccel Big Data Analytics Platform uses the ParAccel MPP data warehouse as a backbone, and blends in Vectorwise and DataRush.
ParAccel's On Demand Integration (ODI) helps tie the pieces together and the ParAccel MPP provides for in-database analytics. This offers data connectivity on the one hand, while minimizing data movement on the other.
Work cut out
Mike Hoskins, formerly CTO at Pervasive and now serving in that role across all of Actian's constituent products and teams, has a lot of work to do. He needs to make some rather disparate IP stacks work together, including not only Pervasive's properties, the ParAccel MPP database and Vectorwise, but the Versant object database, and maybe even the Ingres RDBMS.
That's a tall order, but Hoskins is not underestimating it. And given that he served as the General Manager of Pervasive's Big Data Products and Solutions (in addition to his CTO responsibilities), he's got the head for it. No matter what, industry watchers should keep an eye on Actian, as it attempts to leverage decades of industry experience, expertise and technology solutions in the context of modern Big Data analytics requirements and challenges.