SAP’s In-memory Hana Database Appliance Gets Apped

The old canard the SAP is not an innovator took another hit this week with the partial unveiling of a new set of applications built to run on top of SAP’s Hana in-memory database system. The new apps, due out in a set of rolling releases this year, fill an important gap in SAP’s Hana strategy to date: actual applications that appeal to a line-of-business user, as opposed to Hana’s nerdy technology appeal. Not that there’s anything wrong with that, who doesn’t get all excited about a column-based, in-memory database appliance that can crunch a few hundred million records in a split second or two.

Well, actually, there was one small thing wrong with the last two year’s worth of Hana demos: after the first couple of amazing demos, a certain amount of buzz-fatigue started setting in. After all, once you’ve see the button pressed and the Explorer rows fill instantly with data from a humungous database six or seven times, the gee-whiz factor starts to lose some gee.

So, while SAP didn’t do any comprehensive demos of the new Hana-ready apps, just the prospect of a few good reasons for the line of business buyer to get excited was a refreshing sign of the potential that Hana has to rewrite the good book of data analysis in the very near term.

The Hana apps due to come out this year provide an impressive-looking range of functionality. Below is a list of the application SAP has slated for release this year, mostly in Q3 and Q4.

  • Strategic workforce planning (out now)
  • Sales and operations planning (due Q3)
  • Cash and liquidity management (due Q3)
  • Trade promo management (due Q4)
  • Smart meter analysis (due Q4)
  • Profitability engine (due Q4)
  • Customer revenue performance management (due Q4)
  • Merchandising and assortment management (due Q4)
  • Energy management for utility customers (due Q4)
  • Customer-specific pricing (due Q4)
  • Intelligent Payment Broker (Q4)

While we only saw snippets of a couple of these apps, it’s an impressive list, and adds to SAP’s potential cred as an analytics vendor with some highly competitive functionality. This list clearly comprises eleven in-memory applications that SAP will have by 2012, as compared to all those new in-memory apps the competition has – as in none.

How broadly any of these apps will attack their specific functional domains in release 1.0 requires an understanding of the back story to these apps: They have been developed using an agile methodology that would have been heretical at SAP only a few years ago. When I asked about the 90 days that Vishal Sikka, SAP’s CTO, said it took to develop Strategic Workforce Planning, the answer I got was that SWP (as it is not known internally, I just didn’t want to have to type that mouthful again) and the rest of the Hana apps were being developed without trying to exhaustively deal with the exceptions and special cases that sop up early development resources. Rather, the Hana apps are being developed first to cover the more general-purpose use cases, with the outliers to be dealt with in subsequent releases. So, will SWP 1.0 arrive in the market as the be-all and end-all of workforce planning? Probably not. But it will be the only one in the market that can totally kill the big data problem  behind workforce planning.

Other topics of note in a day crammed with noteworthy topics: Vishal also promised that Business ByDesign would eventually be running on Hana, as would SAP Business One, though no precise timetable was given for when the Hana versions would be released. There was no word on when (not if, IMO) SAP plans to make the decision to move the Business Suite onto Hana as well.

While the full Suite’s time frame for Hana is up in the air, pretty much everything in the SAP portfolio will be Hana-ized in coming years. And those apps will be available as on-demand apps – SAP partner Medidata is already providing its clinical trial management system,running on Hana, as an on-demand application. (And doing some pretty impressive analysis of the metadata being captured by Hana.)

Also noteworthy was the announcement that SAP would be offering to migrate its existing Business Warehouse customers to Hana, a move that would not only give these customers a screamingly fast data warehouse but also significantly lower the DBA costs associated with their BW data base, especially if it’s an Oracle or IBM database. This migration also gives existing BW customers a leg-up in deploying Hana apps, as they will be able to run the new Hana apps on the same Hana engine they are using for BW. Not a bad twofer in Hana’s favor. Actually it’s a threefer:  a faster BW, lower DBA costs, and a Hana engine that can  also be deployed to run Hana apps.

This isn’t just good for SAP’s customers, it’s also pretty bad for its competitors. Especially Oracle, which has bet its business on a stack strategy that SAP is trying to disintermediate with products like Hana. In addition to the possibility that Hana will replace Oracle in the BW market, the fact that it can run on standard, multiprocessor, multicore systems means that the hardware costs for Hana aren’t just cheap, they are decreasingly rapidly as the market for commodity RAM and Intel’s multicore processors just gets better. Vishal noted that a high-end Hana machine his team built last year at a cost of $530,000 was down to $405,000 four months later. I won’t enter the comparison pricing rate race here, as I don’t have enough Hana data to go on, but these prices are likely to put Oracle’s Exadata strategy on the defensive.

In the end, while the details were sparse, there was enough meat on the bone to make it clear that Hana’s future will be targeting the LOB and CEO as much as the IT department: the prospect of advanced, in-memory analytics was always a great idea in search of a critical mass of proof points. By this time next year I expect those proof points to be well ensconced in a new market that SAP is hoping to dominate. The onus is now on SAP’s competition to provide some meaningful competition. The clock is ticking….