Today at a press briefing held simultaneously in Frankfurt, New York and Palo Alto, SAP AG announced that its ubiquitous Business Suite software now runs on its venerable in-memory database, SAP HANA. For a detailed accounting of the news side of the event, read our own Rachel King’s story:
I was at the New York event. The room was full, those in attendance were paying very close attention, and the SAP folks in all three cities were visibly stoked. SAP on HANA is a breakthrough for the German company, for reasons of technology, market share and business models. It may also be a breakthrough for Big Data and analytics amongst traditionally conservative, blue chip companies all around the world.
Analytics, in context
I’ve covered SAP HANA, the German software company’s in-memory database, in some depth before. One of my HANA posts was titled “ .” My point in using that title was that SAP, a company that is focused on business software and which has acquired business intelligence company BusinessObjects followed by database company Sybase, brings a lot of enterprise context with it to the Big Data space. Today’s announcement says that context is no longer in the background, but is on-board, in the in-memory database.
SAP’s Co-Founder, Chairman of the Supervisory Board and Chief Software Advisor, Hasso Plattner, spoke in depth about what it meant to bring SAP Business Suite onto the HANA platform. Essentially, Plattner said that the separation of analytic databases from transactional ones is a hindrance to business success.
Mr. Plattner, tear down this wall
Plattner reasons that transactional/analytical separation is arbitrary. Worse, it introduces unacceptable latency between the conception of business questions and the ability to provide the answers. The latency is unacceptable because the motivation to act on the answer dissipates during the waiting period.
Add to this the growth in business transaction volumes due to online commerce and other factors, and the technology is introducing too much friction. In Plattner’s narrative, disk latency and the separation of OLTP (OnLine Transactional Processing) and OLAP (OnLine Analytical Processing) are smothering business, and business decisions. They both need to be abolished, and the launch of SAP Business Suite on HANA brings about that abolition.
Big Data for big companies
Plattner says SAP thinks business and business analytics both need to run in real time. The subtext is that this shouldn’t be true only for startups and Internet businesses but for the old economy, industrial companies that are SAP’s bread and butter. Hasso spoke from Frankfurt but the action soon shifted to the stage in New York, where SAP’s Head of Sales Rob Enslin was speaking with Derek Dyer, Director of Global SAP Services at Deere and Company (a.k.a. John Deere).
Dyer spoke glowingly of SAP on HANA and explained the company was at times impatient in its desire for the Biz Suite/HANA combination to be readied. In SAP’s press release, Dyer describes the combination as “a game-changer for John Deere” and further explains that the company is standardizing on HANA as a platform for a number of future applications.
When a 175-year old manufacturer of heavy machinery gushes about in-memory databases and Big Data analytics, you know they’re both going mainstream.
While SAP was already in the database business with its acquisition of Sybase, running its Business Suite on its own database is the company’s own real game-changer. SAP implementations today run on the major relational databases (Microsoft’s SQL Server, Oracle and IBM’s DB2). And while SAP was careful to point out that the new bits will continue to run on those platforms, HANA gives SAP the ability to own the entire software stack above the operating system layer.
So in one fell swoop, putting Business Suite on HANA gives SAP a big competitive edge in the Enterprise Resource Planning (ERP) world, a major presence in the Enterprise analytics world, and greater control of its own destiny. Not bad for a company that many were dismissing as a dinosaur just a few years ago.
Will customers get the religion?
Plattner said that in some tests, SAP on HANA was 250,000 times faster than on conventional relational databases. What remains to be seen, though, is whether a significant number of SAP customers feel they need their ERP thusly accelerated and are willing to pay for the privilege. There is only so much memory that can go on a given server, after all. And scaling out the kind of iron that runs HANA is a bit different from adding a few more blades with commodity disk drives to a Hadoop cluster.