SAP HANA, the company’s in-memory appliance, took top billing and once again dominated the keynote sessions at SAP’s premier business event, SAPPHIRE Now. This year in Madrid the company announced, amongst other enhancements, a version of HANA that runs SAP NetWeaver Business Warehouse (SAP BW), enabling users to tap into the high speed loading and query performance of HANA for near real-time analysis of large data sets and instantaneous insights. This announcement sees SAP evolving and widening the use case for HANA (from its use as a high speed ERP operational data mart environment) at the same time as providing an opportunity to seed its in-memory technology across other parts of its business analytics customer base.
BW is evolving to a fully in-memory enabled data warehousing solution on top of HANA
BW is SAP’s enterprise data warehouse system that has been deployed by over 12,000 customers. It packages together ETL, a data warehouse including a star schema (known as an Infocube), and reporting, planning and analytic capabilities. Although a successful product for SAP, the company has, for some time now, looked towards in-memory technology to help improve the processing and analysis performance of its data warehousing offering. It was therefore only a matter of time before SAP announced a tighter interplay between HANA and BW to enable customers to engage more cost-effectively in faster and near real time planning, reporting and analysis of BW data.
In practice, running on top of HANA means some of BW’s core functions such as the calculation and planning engine, analytic indexes and data store objects can move in memory. Loading, processing and analysing data in memory means faster reporting, shrinking batch windows, the ability to build aggregates on-the-fly as well as removing the need to build some Infocubes, (although not all) which helps reduce DBA maintenance and development effort as well as storage space. However, one of the biggest benefits the company is advocating is that this configuration will no longer require a relational database layer such as Oracle or IBM’s DB2 for persisting BW’s Infocubes. By running BW on top of HANA, data traditionally stored in a database can now be moved and stored in an in-memory database. The idea here is not only to make things run lightening fast but also to reduce the layers of infrastructure needed to run your SAP BW environment and thus reduce or at least simplify system administration and operating costs.
SAP has Oracle in its sights
This is clearly a competitive shot at Oracle in particular; the company (much to SAP’s frustration) has a significant number of customers, including BW customers, running on its database. Not only that but Oracle is also advocating Exadata Database Machine as a high performance engine for running SAP applications. By making HANA an attractive option from both a BW performance and cost perspective (through potential database licensing and maintenance savings) SAP hopes to tempt many of these customers to make the switch and migrate to HANA.
That said, SAP’s HANA plans go much further than BW; the longer term aim is to leverage it as a foundational platform for running the company’s existing application portfolio as well as hosting a range of new in-memory enabled apps. In convincing many of its loyal BW customers to move to HANA, SAP is hoping to put them on the right path, so that when – or if – they choose to run their business applications on HANA, they are in a stronger and more informed position to do so. But at the same time this also represents a significant test for HANA and SAP’s in-memory strategy. If the company is successful in migrating many of BW’s 12,000 customers then it will most certainly pave the way for future development and HANA adoption across other parts of SAP’s portfolio. SAP has invested a lot in HANA and it needs to make it a success.
A smooth migration process is key to success
Part of this success will depend on ensuring a pain-free or smooth transition from running BW on a relational database to HANA – especially when migrating from what is perceived as a reliable database to something that is still relatively unproven may prove to be a significant challenge and risk for some. To allay these concerns SAP has been bullish in its claims of “innovation without disruption” and the relative ease and speed with which customers can migrate to HANA from a traditional BW deployment. We will of course be interested to hear from any customers who can either refute or corroborate these claims. You can reach me at firstname.lastname@example.org.
Notwithstanding these migration concerns, one of the biggest questions raised about BW on HANA remains its relationship with Business Warehouse Accelerator (BWA). As mentioned previously this isn’t the first time SAP has leveraged in-memory to boost the performance of BW. In fact, BWA can, in many ways, be seen as a predecessor to HANA on BW in that it loads parts of the application in-memory and leverages BWA indexes for faster response times and performance. Given the obvious overlaps in functionality, being clear about the use cases and future road maps of each product must remain a priority for SAP. When questioned at SAPPHIRE, SAP told us that existing customers can continue to use and invest in BWA but also have the option to move to HANA if they also wish to gain the additional benefits of improved ETL performance and simplified IT infrastructure. Although SAP proposes to enhance and maintain both products, we do believe that over time HANA is likely to supplant BWA as an acceleration tool especially if adoption of BW on HANA proves to be a big success. We looking forward to tracking this story with interest.