SAS said the approach is about to be tested with a handful of early adopter customers, with candidates including banks doing credit-risk analysis, insurance companies doing fraud detection and campaign targeting, and retailers interested in price optimization.
It's no surprise that SAS would favor independents Teradata and EMC with an early exclusive, but Wilkie did not rule out opening up HPA to other partners down the road.
IBM could not be reached in time to comment on the news.
With yesterday's announcement by SAS, EMC Greenplum is taking a big step forward in analytics. And bolstering its appliance lineup, ECM added its own announcement of two new EMC Greenplum appliances and a data-integration accelerator module.
EMC is following in Teradata's footsteps with the new appliances and the integration module, but it's promising better price-for-performance stats.
The new High Capacity ECM Data Computing Appliance (DCA) is aimed at analysis of petabyte-scale data stores. Telcos often require high-scale deployments in which they query vast records in search of particular call data records (CDRs) for compliance reasons. Cost usually trumps speed as the top concern when meeting regulatory or archival requirements.
The High Performance EMC DCA uses solid-state drive (SSD) technology to ensure fastest-possible data loading and querying for time-sensitive analysis. SSDs are said to deliver 150 times faster data input/output per second than achievable with conventional spinning disk drives. Speed comes into play in near-real-time decision-support scenarios in which you need to make decision while a customer is still on the phone or surfing your Website.
Teradata took on high-capacity and high-performance requirements 18 months ago when it introduced its Extreme Data Appliance (akin to EMC's new high-capacity appliance) and Extreme Performance Appliance (akin to EMC's new high-performance appliance).
Cost and capacity are constantly moving targets, so I'll refrain from a hasty analysis. EMC has published capacity and performance stats here.
EMC's new Data Integration Accelerator is an add-on module that moves data-integration to the ECM Greenplum appliance platform, taking advantage of its parallel processing and scalability. The module dedicates processing nodes to data-staging, and it's aimed at companies seeking near-real-time data loading and analysis by way of shorter batch loads or even continuous micro-batch loading.
EMC's Data Integration Accelerator is akin to an option Aster Data previously introduced. Monash of Monash Research described EMC's two new appliances and the integration option as "a smart bit of catch up."
The stakes are high in the analytics arena these days, and nowhere higher than in the big-data and predictive analytics domains. That has led to aggressive development and a lot of visionary announcements. That HPA won't be generally available for at least six months isn't surprising. SAS's initial in-database work with Teradata took more than a year to reach the market, but it's now a standard approach.
So it's a good bet that modeling work -- or at least certain types of SAS modeling work -- is about to move into the fast lane -- at least with two favored partners.