Teradata is adding an appliance option for its Aster analytic database, giving it another potential weapon against rivals such as Oracle’s Exadata, EMC’s Greenplum and SAP’s HANA.
The product, announced Thursday at the Strata conference in New York, will be built with the same infrastructure that underpins Teradata’s range of data warehousing appliances. It joins the existing software-only and cloud deployment options for Aster’s database, which was acquired through the purchase of startup Aster Data Systems in March.
The appliance is expected to be released in the first quarter of next year, with pricing yet to be determined, said Randy Lea, vice president for the Aster Data Center of Innovation at Teradata.
Aster has made much of its support for the MapReduce programming framework for large-scale processing of data, particularly “big data” types such as that generated by Web logs and sensors. Its SQL-MapReduce toolset allows users to invoke MapReduce functions from within business intelligence tools or with standard SQL.
Aster Database 5.0, which was also announced Thursday, adds a number of improvements to that framework, including “pre-built MapReduce modules for behavioral clickstream interpretation, marketing attribution, decision tree analysis and other analysis,” the company said in a statement.
Other new features focus on workload management and SQL performance.
The 5.0 release is scheduled to arrive early next year along with the appliance, Teradata said.
While both Aster and Teradata’s own database are aimed at analytic workloads, they have differing and complementary strengths, according to analyst Curt Monash of Monash Research.
“Aster is best used for two things,” he said. “First, you can do complex analytics on a research basis. Second, after your research succeeds, you can massage data and derive conclusions that are fed into a more operational, even if analytic, database.”
In turn, “Teradata is a great system for running operational analytic databases. In many cases, it’s the single best choice.”
“Teradata is good for a broad variety of analytic tasks, and most especially for use cases that combine several of them,” Monash added.