SAP plans to release its High-Performance Analytic Appliance (HANA) in November, roughly six months after announcing the product.
The HANA machines employ in-memory technology SAP has been developing for some time. In-memory processing pushes data into RAM, giving a performance boost over reading it off of disks, proponents say.
IBM and Hewlett-Packard were already set to supply hardware optimized for HANA appliances, but SAP is set to announce Tuesday that it is also developing “co-innovation” programs with Cisco, Fujitsu and Intel.
Overall, SAP is positioning HANA as a faster, better way to process data from business applications in real time.
SAP is also readying a series of specialized applications, including one that utilities can use to analyze data from smart meters, said CTO Vishal Sikka in an interview Monday.
SAP is no doubt hoping customers will use its own Business Objects BI (business intelligence) software on top of HANA for reporting and visualizations.
But HANA will be compatible with any other BI application that supports common query languages like SQL and MDX, said Sikka, who will deliver a keynote at SAP’s Tech Ed conference in Las Vegas on Tuesday.
Currently, HANA is supported for use only with Intel chips, Sikka said. He did not rule out future support for AMD silicon, but indicated SAP has no intentions to build out support for Oracle’s SPARC chips, which it acquired through the purchase of Sun Microsystems. Only a relative few SAP customers are running their applications on SPARC-based servers, he said.
Oracle has its own entry into the high-performance appliance wars in the form of Exadata. The company is taking a far different approach than SAP, touting the benefits of a tightly integrated set of components and software from a single vendor.
“It is possible to have strong support for [multiple hardware vendors] and still have unbelievable performance,” Sikka said. Oracle is presenting a “false choice” between heterogeneity on one hand, and performance on the other, he added.
SAP’s stance is pragmatic as well, given it has many customers already running IBM or HP iron, said Forrester Research analyst James Kobielus.
At the same time, IBM has many products that compete with HANA, and recently raised its bet in analytics significantly by moving to purchase data warehousing vendor Netezza.
But that’s no reason to believe IBM’s efforts to optimize its systems for HANA will be lackluster, Kobielus said. “SAP has many, many customers that run SAP on IBM hardware. IBM does not regard those customers as being less deserving of high performance just because they haven’t licensed the entire IBM stack,” he said.
The first HANA appliances will be released to customers on Nov. 30. However, as is SAP’s custom for new products, an initial “ramp-up” period will ensue before a broader rollout, Sikka said.
Pricing was not available at press time.
Chris Kanaracus covers enterprise software and general technology breaking news for The IDG News Service. Chris’s e-mail address is Chris_Kanaracus@idg.com