HANA On Power Marches Toward GA
March 4, 2015 Alex Woodie
SAP is now inviting its Power customers to participate in the final beta program to harden the HANA-on-Power code before it’s released to production. IBM Power Systems and IBM i customers are said to be “champing at the bit” about the prospect of using a single in-memory database to power both transactional and analytic applications in a single server.
SAP officially unveiled its plans to get HANA running on Power last June at its Sapphire conference in Orlando. Over the summer, less than 10 joint IBM-SAP customers participated in a tightly controlled “test and evaluation” program that involved the first whack at running HANA on Power Linux. At its TechEd conference in October in Las Vegas, SAP announced its “ramp up” program that opened the beta a little wider.
Last week, SAP began accepting requests from its IBM Power Systems customers to enroll in an even wider beta to test the final HANA-on-Power release candidates before the product is brought to general availability. The customers can get more information by going to SAP’s Ramp-Up webpage and signing in with their SAP user ID (SID).
Joint IBM-SAP customers are eager to get started with the HANA on Power program, says Alex Cabanes, the program director of ISV business development and solution marketing in IBM’s Systems and Technology Group.
“The ramp up program has really been exciting because customers have been champing at the bit to get in,” Cabanes says. “It is awesome. I have to defer to the SAP team because it’s their product, but we’re definitely excited about it and seeing a lot of interest about it in the marketplace.”
Some IBM-SAP customers have been so gung ho about the program that Cabanes has had to reel them back in. “We’re almost having to tell customers ‘Hold on a second.’ The demand has just been overwhelming,” he tells IT Jungle.
All the hub-bub is around HANA, an in-memory, column-oriented, relational database developed under the guidance of SAP co-founder Hasso Plattner (some call it “HAsso’s New Architecture”). The first iteration of HANA–available only for servers running Intel X86 processors of course–was focused primarily on analytic workloads. But with the unveiling of S/4 HANA last month, SAP is now offering a version of its flagship Business Suite that runs atop HANA.
During the launch of S/4 HANA last month, SAP talked glowingly about “reducing the data footprint” and “reintegrating ERP, CRM, SRM, SCM, and PLM.” The proliferation of business systems has created too much operational diversity in the data center and needs to be de-fragged, the Germans say.
By moving historical data out of the database and keeping only relevant data running within an in-memory database, applications run from 3x to 5x faster, SAP says. That gives customers the choice of running their entire business on a system that’s a fraction of its previous size. And since operational analytics is baked right in, there’s no need for separate data warehouses.
Now that we’re only months away from seeing a version of HANA that runs on IBM’s Power-based systems, executives at the two tech giants, as well as their customers, are getting downright giddy about the prospect of putting a Power-based HANA system up against HANA running on Intel X86.
“All the attributes that have been historically associated with Power Systems around reliability, availably, security, and scalability–all of those strengths basically now translate to the HANA environment,” Cabanes says. “The customers I talk to . . . have made the strategic direction to go to HANA. They’re the leading edge, pushing the envelope. They’re looking for a more robust environment, because they’re betting their company on this.”
Cabanes was mum about the prospect of IBM and SAP developing a reference architecture for HANA on Power, or about going to market with a pre-configured “solution edition” approach. But clearly both of those will be in the works (if they’re not already). And chances are very good that we’ll start hearing about how HANA-on-Power will look in the real world at the Sapphire conference taking place May 5-7.
Cabanes provided a glimpse of what the marketing around HANA on Power will look like. “It comes down to three buckets that we’re focused on,” he says.
The first is “run simple.” “[HANA on Power] reduces the complexity of having a scale-out environment by allowing you to consolidate all of the pieces in a single system,” Cabanes says. “It reduces the potential points of failure. It reduces the complexity associated with managing the system and really being able to run your business on it.”
The second bucket is “run fast,” which comes down to the superior threads per core and increased memory bandwidth of the Power 8 architecture.
The third bucket is “run smart.” “When you look at the Power architecture and the predictive failure alerts and basically the self-monitoring of the hardware [and compare it to] some of the implementations of the Intel X86 architecture, the Intel X86 architecture is much more reliant on heartbeats,” Cabanes says. “If a system skips a beat– if the system crashes–then you have to go into major recovery.”
In the meantime, there is still code to test.