In-Depth
Teradata is Thinking Big
With its forthcoming Teradata v14 release, Teradata is going columnar, and that's just for starters.
- By Stephen Swoyer
- 08/16/2011
Refreshes of Teradata Corp.'s flagship database platform are almost never humdrum, but the forthcoming Teradata v14 release might be one of the company's biggest -- or most feature-packed -- releases yet.
For starters, Teradata plans to deliver its long-awaited columnar capability, Teradata Columnar. Unlike some of its competitors, Teradata isn't delivering a bolt-on or complementary columnar capability, officials argue. Teradata Columnar, they claim, is embedded in the Teradata 14 database itself.
"What is really differentiated about what we're doing is we're embedding it inside the Teradata database ... we're not just delivering an analytic environment that is specific for analytics that are wide tables," says Chris Twogood, vice president of product management with Teradata.
"It basically becomes a feature of the Teradata database, so users can determine [if] it's a column or if it's [a] standard [row] that they want," he continues. "What we did is [we took] our row partition and turned it on its side and [made it] a vertical partition. So [Teradata Columnar] is a natural extension of capabilities that we already had."
Teradata plans to unveil Teradata 14 at its forthcoming Partners user conference, which is slated for San Diego in early October.
Also at Partners, Teradata plans to unveil Unity, a much anticipated offering that supports user and query routing, as well as data and database synchronization capabilities between and among Teradata systems.
Unity incorporates technology (the former "Gridscale") that Teradata acquired when it purchased database virtualization specialist xkoto last summer.
Unity is Teradata's answer to an increasingly common problem: namely, customers tend to have multiple analytic systems, each of which is tuned for different data and workload requirements.
Customers can configure Unity to redirect queries to specific systems based on data or workload requirements. "It fundamentally enables a query to go into Teradata Unity ... [where], through intelligent query routing, Unity itself will say where's the best place to enable the query to execute," Twogood explains. "It also supports synchronization, so any changes within those data structures can synchronize across all of those environments."
Also in the Partners timeframe, Teradata plans to release a connector that facilitates high-speed, bi-directional integration between its own environment and Aster Data Systems Inc.'s nCluster environment. Teradata purchased Aster in February of this year.
Twogood and Teradata are still relatively mum concerning Teradata's plans for the former KickFire Inc., an analytic database start-up that it purchased in August of 2012. "[W]e continue to be interested in that technology for streaming," says Twogood, referring to KickFire's SQL-on-a-chip FPGA technology.
The primary impetus was to acquire KickFire's brain trust, he explains.
"Mostly it was the people we bought [i.e., as the reason for the acquisition]. As we looked under the hood of the technology, we found that it was really good but not mature enough. So we're toying with different places to use it, but we haven't defined specifically where it's going to," he concludes.