SAP’s HANA and Sybase IQ are separate but complementary

SAP is en route to having two column-based, in-memory databases in its portfolio, but far from seeing them as competing or conflicting products, it regards them as complementary. Ovum investigates its claims and concludes that the databases address different markets, and will even be able to work together for melding data from SAP and non-SAP environments.

HANA is now front and center of SAP’s agenda

The recent combined Sapphire and TechEd conference inMadridhad in-memory computing as one of its main themes, with SAP highlighting to customers and the media the in-memory analytical capabilities of its HANA software.

When it acquired database vendor Sybase in May 2010, the German software giant got a separate line of development of in-memory database (IMDB) technology. Sybase’s RAP platform, which is marketed heavily into the capital markets, comprises three “engines”: an in-memory version of its flagship ASE relational database, the column-based IQ database, and a complex event processing (CEP) engine.

There is now an in-memory, column-based database at the heart of the HANA platform, and Ovum is reliably informed that an in-memory version of IQ is also in development. This raises the question of whether in the longer term SAP plans to maintain HANA and IQ as separate code bases.

HANA and IQ won’t compete, even when IQ goes in-memory

For the moment at least, the answer is yes. HANA is sold into the existing SAP customer base either to turbo-charge the company’s core Business Suite of products, or as announced at the conference, to replace a conventional disk-based relational database underpinning a data warehouse, in particular the Business Warehouse from SAP itself. There is also work under way with EMC to enable HANA to underpin EMC’s Greenplum data warehouse. In terms of the financial sector, HANA is sold to the many retail banks that use Business Suite, either as their core banking system, or for other purposes such as a general ledger.

By contrast, Sybase RAP is aimed squarely at the capital markets where its CEP engine can handle the intense flows of streamed market data, while IQ can hold, say, the last few days of market data and ASE can hold large amounts of historical data for comparison purposes.

There are, however, signs of convergence between the development tracks. While IQ is going in-memory, HANA will soon begin to span the retail and investment banking worlds when SAP delivers applications, such as a realtime risk-analysis module, to sit on top of the platform. And in the longer term, there is actually the potential for the two to work together.

IQ could become an on-ramp for HANA

One of the big differences between IQ and HANA is that the Sybase product is heterogeneous in its capabilities, and can handle structured and unstructured data from any source. HANA, by contrast, is largely for SAP environments, although developments such as the work with Greenplum will extend its functionality. There is, however, the possibility that IQ might act as a kind of near-line storage for non-SAP data, consolidating, normalizing, and optimizing it all in preparation for injection into HANA to be processed alongside data pulled from SAP systems.

Whenever a big, diversified company like SAP acquires a smaller one with a range of assets like Sybase, there is always the “embarrassment of riches” scenario in which the acquirer may have a project under way that to some degree overlaps with something that the acquired company was also working on. In SAP’s case, it appears to be handling the perceived overlap between HANA and IQ (both column-based and, soon, both in-memory), recognizing that each product has a different role to play, that each has its own strengths, and that in some circumstances the two can actually work together in a complementary way.