By Charles King, Pund-IT, Inc. June 11, 2014
Last week’s SAP Sapphire NOW user and partner conference in Orlando was something of a love fest for the company’s HANA in-memory database technology. That was hardly surprising, since SAP has made it clear that HANA represents far more than just an innovative solution for a variety of big data, analytics and business intelligence workloads. In fact, the HANA platform, along with an evolving number of complementary solutions, including the new cloud-based PaaS and subscription offerings SAP announced in March and April represent the elemental future for the company and its customers and partners.
That’s a significant commitment by most any measure, so we were intrigued by a particular element in one of the press releases SAP published during the conference; that the company and IBM have begun a testing and evaluation (T&E) program for HANA running on IBM’s POWER7+ and POWER8-based systems. Specifically, the program allows for select IBM and SAP clients to deploy HANA solutions on Power Systems in SuSE Linux-based environments. Why is this of interest? Because until now, SAP HANA deployments have all been in Intel x86-based systems and appliances.
In fact, the list of hardware vendors leveraging HANA reads like a who’s who of Intel-based systems players. Moreover, this clear focus on x86 has provided SAP a notable weapon in the increasingly competitive and contentious big data market. Most other in-memory platforms consist largely or entirely of a single vendor’s proprietary technologies, including Oracle’s Exalytics and IBM’s Blu Acceleration. In contrast, HANA is open to any vendor that can meet SAP’s stringent certification process, enabling numerous players to effectively enter and compete in big data market opportunities.
Why HANA on POWER?
Given those points, why would SAP and IBM decide to cooperate on certifying HANA for Power Systems? Mainly for strategic, go-to-market reasons. Though big data certainly garners more than its share of attention and hype, demand for associated solutions is still in early days, especially beyond deep-pocketed enterprises and analytics/BI specialists. As SAP continues to develop the demand for HANA, it behooves the company to pursue any and all serious partnerships, which certainly includes IBM.
Similarly, IBM is working hard to explore new opportunities for its Power Systems and new POWER8 microprocessors. That’s partly a matter of self-preservation. Depending on whose market data you read, IBM owns near or over half of the market for traditional Unix systems, but given the continuing implosion of Unix solution sales (contracting at near 10% per year), cultivating other options is absolutely critical. That’s part of the genesis of the OpenPOWER Foundation, launched in 2013 by IBM and partners, including Google, NVIDIA and Tyan, and that has grown to over 30 members since then. In many real ways, the new HANA/Power T&E program fully complements and should significantly extend the OpenPOWER effort.
Plus, the two companies have a long history of friendly, innovative collaboration. IBM and SAP both focus largely on global enterprises and industries, and the pair shares numerous common customers. But they have also worked in concert to develop notable new solutions, including HANA-based offerings. In fact, IBM’s System x has long been the reference architecture for SAP HANA, partly because IBM’s X-Architecture technologies can support significantly greater memory capacity (in the latest X6 generation, up to 3X more memory) and better performance than competitors’ solutions.
The Power Difference
This brings up a related practical point for why the HANA on Power Systems T&E program is important for IBM: Without it, the company’s planned upcoming sale of its System x organization to Lenovo will leave IBM without a native solution for HANA. If SAP is successful in persuading customers to standardize its applications on the HANA platforms, IBM would be essentially locked out from participating – not an appetizing turn of events.
Given that, an interesting question to consider is whether IBM Power can or will offer SAP and its HANA customers anything they don’t already have or can’t get with other solutions. On paper, the match-up looks pretty good for IBM. In a recent blog post, Terri Virnig, VP of IBM Power Systems Ecosystem, stated that the company’s latest Power solutions “Will enable SAP HANA to run more queries in parallel faster, across multiple cores with more threads per core than possible on x86-based servers. HANA on Power will also gain increased memory bandwidth and faster I/O to ingest, move and access large volumes of data faster.”
What that will actually mean for IBM sales is another thing entirely. Virnig is understandably optimistic, noting that those and other Power qualities “Make a compelling argument for customers currently running HANA solutions on x86-based platforms to run HANA on Power Systems.” But the decision to effectively abandon one management-sanctioned platform for what amounts to a new solution is not one than many (or any) companies take lightly. In order to justify such a decision, the provably achievable benefits need to be considerable.
The HANA on Power Systems T&E program should go a long way toward demonstrating the kind of benefits customers can expect from integrated SAP/IBM solutions. If Power delivers the kind of HANA-based performance that IBM claims, the news should be very good for the company.
But it will also inject additional choice and competition into the in-memory analytics/big data market, eventually resulting in significant price/performance benefits attractive to both existing customers and new clientele. We also believe the effort could help to validate analytics as a sort of lingua franca for next generation IT; an implicit point in the efforts of many big data-focused vendors.
The fact is that as data and data sources continue to accumulate and accelerate, it will be increasingly difficult for organizations to realize the full benefits of their information assets and investments. SAP’s strategy around HANA has succeeded by emphasizing how a unified in-memory platform can deliver the goods in both analytics and common line-of-business applications.
If IBM’s Power Systems can deliver the kind of enhanced performance and results the company claims, that’s all to the good. But just as important will be the increasing competition that viable Power-based HANA solutions should inspire in the marketplace. The fact of the matter is that big data belongs everywhere, not just in enterprise or analytics ivory towers. If SAP and IBM’s HANA-on-Power Systems effort succeeds, the two companies could go a long way to achieving just that.
© 2014 Pund-IT, Inc. All rights reserved.