During the Run Better Tour in Boston last Wednesday SAP announced an Applications Roadmap for its In-Memory Technology HANA (High Performance Analytical Appliance). Invited as a blogger and SAP Mentor, I had the opportunity to take part in meetings with SAP board member Vishal Sikka and CIO Oliver Bussmann.
Business Warehouse on HANA
BW applications on HANA run at supercharged speeds, taking only a fraction of their old-skool database counterparts. For Vishal Sikka this is the next evolutionary step to replace a common database for Business Warehouse (BW). End goal here is to also use HANA one day as the transactional database of an ERP system.
At this point in time in-memory is still associated with a hefty price tag. A critical mass of analytical data is necessary to justify the technology (currently a server with 10 blades and 32 processors each costs roughly $500,000). However it is already becoming clear that hardware costs for this are coming down rapidly. Vishal Sikka explained that one of SAP’s HANA systems hardware costs dropped by 25% in the space of only 4 months. Lower price segments for HANA-capable systems are definitely around the corner.
The Roadmap for new, on HANA based applications comprises
- Sales and Operations Planning
- Cash and Liquidity Management
- Trade Promo Management
- Smart Meter Analysis
- Profitability Engine
- Customer Revenue Performance Management
- Merchandising and assortment management
- Energy management for utility customers
- Customer-specific pricing
and is due to be delivered in Q3 and Q4 2011. Whether a list such as this really resembles a roadmap is arguable, but the attack with which SAP is aiming to deliver these solutions is inspiring.
A real reality
Product demos of Strategic Workforce Planning and Smart Meter Analysis demonstrated HANA’s impressive response times. Queries that used to take hours are now processed within the split of a second. However, I would like to see more complex scenarios during those demos. As impressive as the speed of re-pricing during the Smart Meter Analysis demo was, I think a real world customer repricing would be more complex and require specific parameters. Now that interested customers are beginning to grasp the potential of in-memory, it becomes even more important to show the “real world” capability and use case of it.
Moreover it will also be important to mitigate the danger of existing Netweaver customers feeling left aside amidst all this high speed innovation. It will be paramount for SAP to “pick up” those customers. The fact that migration paths towards HANA for existing IBM and Oracle database customers will be offered later in 2011 is a very encouraging sign. Walldorf has to walk a tightrope of traditional business and innovative technologies whilst being non-disruptive. Not an easy feat.
Process changes through In-Memory Technology
One aspect that deserves more exploration in my view are changes to work and business processes through in-memory technology. How immediate does In-Memory need to be? Especially “freshly” created transactional data -like sales orders, for example- often undergo frequent changes after creation. How reliable are decisions based on data such as this? In addition, what can users do to make data more stable and reliable? Is there a potential for more intelligent software to guide this?
Dr. Sikka promises: “Every SAP product will go through a transformation.”. Since 2009 innovation has been extremely enforced at SAP and new methodologies such as Agile Development have been introduced. The software concern is undergoing the biggest change in its history of almost 40 years and the speed of new developments is breathtaking: Strategic Workforce Planning has been developed within 90 days, for example. Whether a concept of leaner, simpler applications with faster time-to-market is the winning formula remains to be seen, however early reactions seem to prove SAP right. Vishal Sikka has plans for an ERP on HANA, but does not make any promises or commit to timescales here.
From a sidecar to a limousine
In the long run HANA is supposed to – as announced in 2009 by SAP veteran Hasso Plattner – replace the database for transactional data. Up to now HANA works like a “sidecar”, replicating transactional data. This would be the final and biggest step in this evolution. It would also open a door to SAP that has been regarded as tightly shut for years: Regaining part of the enterprise database market share, as most SAP customers use Oracle and IBM for their ERP database systems.
note: the german version of this post below can be found on my Silicon.de blog.