Teradata has been in the business intelligence (BI) business for some time, making purpose-built appliances for data warehousing applications. SAS has been in the BI business for decades as well, historically with industry-leading statistical software, but more recently with predictive analytics technology as the flagship component of a full BI stack.
Now the two companies have teamed up, with the introduction of the Teradata 700 appliance for SAS High-Performance Analytics. The appliance runs a version of SAS' analytics technology called SAS High-Performance Analytics for Teradata. Together, the appliance and the software bring together many of the technologies and architectures we've discussed in this blog:
- Appliance technology, such as is used in Massively Parallel Processing (MPP) products
- In-memory database technology
- Parallel processing of analytics tasks, which bears certain affinity to both Hadoop/MapReduce and MPP
The SAS High-Performance Analytics software includes SAS data exploration, model development and model deployment software. According to Teradata, in the case of one of its existing customers, the combination of the 700 appliance's finely-optimized hardware, the in-memory technology and the parallel processing approach resulted in "improving modeling and complex query performance from a week down to less than 2 minutes."
I'm guessing that not every customer will benefit to that degree, but we don't really know yet. This is brand-new technology, being announced concurrently with the posting of this article. There's further information on Teradata's Web site and I intend to dig in and learn more.
One thing seems certain, though: we will continue to see the convergence of BI and Big Data, along with numerous combinations and permutations of data mining/predictive analytics, in-memory database technology, column store databases, NoSQL, distributed processing and appliances designed around it. Let's just keep an eye on which combinations end up being the most compelling. Perhaps this Teradata-SAS tie-up will be on our list.