Catching up with Actian ParAccel
What Actian's acquisition of ParAccel means for BI professionals.
- By Stephen Swoyer
- August 27, 2013
Three months after its acquisition by Actian Inc., ParAccel is ready to answer questions, but not about its product or technology road map -- at least not just yet.
It's still too soon, stresses John Santaferraro, vice president of marketing for Actian's new ParAccel Platform Group. "The road map is not there. However, we're happy to talk about the obvious connection points [between Actian's products] and promise that we will be back in the next couple of months with an actual road map," Santaferrao told BI This Week.
When Actian kicked off its bid to acquire ParAccel this April, officials were vague about just how ParAccel, which marketed a massively parallel processing (MPP) database platform, was to fit into Actian's product stack. For one thing, Actian marketed an analytic database offering of its own in VectorWise, a non-MPP symmetric multiprocessing (SMP) platform.
Actian's acquisition gambit suggested a tantalizing possibility, however. If the SMP bits from VectorWise could be combined with the MPP bits from ParAccel, Actian would lay claim to a database engine that could scale both vertically (that is, across multiple processors, on a single system) and horizontally (across multiple nodes in an MPP cluster). In the data warehouse (DW) space, most players focus on one or the other -- much as Actian and ParAccel, as separate companies, once did. Actian officials wouldn't discuss specifics, however.
"The real end-game is to offer customers a smooth path with a single vendor and a single infrastructure vision where they can grow from their current needs to their future needs," Actian CTO Mike Hoskins told BI This Week in an April interview. "We're not revealing anything. We're fast and furiously working to reach a more scalable solution. At the very least, we have loose coupling [between the two platforms] that works out of the box."
Santaferraro, interviewed at last month's Pacific Northwest BI Summit, offered to clear up the confusion. "If we can make that [MPP] ParAccel engine core-aware, we can push analytic [workloads] out [across an MPP cluster] and distribute [it] across 20,000 cores if we want to," he said.
"That's our focus [in the ParAccel Platform Group]. Today, we have the capability of doing anything from a single app running high-performance analytics to supporting an entire analytic ecosystem on the high end -- everything from a single-node SMP system to massively parallel processing. What we want to do is combine that so we have core-aware [i.e., SMP] and massively parallel processing in the same system."
Actian's acquisition of the former Pervasive gives it still another intriguing platform, especially from a scalability perspective. Pervasive specialized in scalable data integration (DI) products, starting with DataRush, its massively parallel ETL tool.
Thanks to a recent rebranding, DataRush is now dubbed the "ParAccel DataFlow Engine." Pervasive's other Rush-branded products -- RushAnalytics, RushLoader for Hadoop, and RushAccelerator for KNIME (an open source analytics platform) -- have also been brought under the ParAccel umbrella brand.
The plan, says Santaferraro, is to offer point-and-click data selection, preparation, extraction, and movement -- all within the ParAccel environment, all from a single console. Pervasive purported to do something like this with RushAnalyzer, a tool designed to simplify the selection and preparation of analytic data sets.
By shifting the Pervasive DI products into the ParAccel Platform Group, Santaferraro says, Actian expects to focus on automating the flow of data, from provisioning to ingestion.
"We want to offer the ability to do full data provisioning and data flow automation using the Rush technology. If you think of data in motion, data as a flow, instead of [data as] something that's always being persisted, we have the ability also to connect these data flows, from point to point," he argued.
"This means connecting to the data, preparing and enriching it, applying operators to do data transformations, do data profiling, to do things like match and merge, to do a number of preparatory things. We can do that in-flight using what is now the ParAccel DataFlow Engine -- or we can [push] that in[to] the database."