Revisiting SAP’s Big Data Business

Ask 10 people to define the term “Big Data” and you’ll get 10 different answers. That in turn makes it difficult to determine the total Big Data market size, since intelligent, well-meaning people can and usually do disagree over which vendors and products to include and which to leave out.

Of course, we didn’t let that stop us.

Since the moment Wikibon published its Big Data Market Size and Vendor Revenues report in mid-February, we’ve been hearing from many of the vendors in the rankings. Some were happy with their position (read: we may have overestimated their revenue!), while some were not. Our position is we want to hear any and all feedback, and when a vendor (or anyone else for that matter) makes a persuasive argument for amending the rankings, we will do just that. Think of Wikibon research as living documents.

To that end, after a number of discussions with the vendor, input from the Wikibon community as well as several hours of internal debate, Wikibon has increased its estimate of SAP’s Big Data revenue from $10 million to 85$ million.

As way of background, Wikibon originally determined that HANA, SAP’s new in-memory database, while an interesting new approach for the vendor, did not fall under the Big Data umbrella because most HANA deployments did not handle large data volumes (less than 10 terabytes in most cases) and the data concerned was largely traditional, structured data. As for Sybase, whose IQ analytic database is columnar and runs on commodity machines, we simply didn’t get much feedback on many large-scale deployments. The $10 million figure was mainly related to the use of the Business Objects portfolio of business intelligence and data integration tools in a handful of Big Data projects.

As for HANA, we have determined that it does indeed deserve inclusion due to its non-traditional (purely in-memory, parallel processing) approach and its real-time predictive analytic capabilities in deployments where data volumes exceed what are traditionally considered “large” by SAP standards. That means tens of terabytes or more. Sybase, meanwhile, restructured its IQ databases to a shared-everything, massively parallel processing architecture with release 15.3 last summer, and added Hadoop integration capabilities shortly thereafter with release 15.4. Like HANA, we don’t consider most Sybase deployments to qualify as Big Data, but in those cases where data volumes and variety merit we have included the associated revenue.

The upshot of all this is that in addition to the $10 million in Business Objects-related revenue, we have added $52 million in HANA revenue (out of a total $212 million HANA revenue for 2011) and $25 million in Sybase revenue (out of an approximate total of $250 million Sybase database revenue – this figure excludes mobile-related revenue- for 2011) to SAP’s total Big Data revenue. SAP’s total Big Data revenue now stands at $85 million and we have updated the report to reflect that.

SAP is in a much better position to help its customers make the transition to the world of Big Data than many give it credit for (including, until now, us.) Between Sybase IQ for large-scale data warehousing, Sybase RAP and CEP for streaming analytic workloads, and HANA to support lightening-fast analytic applications, SAP has most of the raw materials in place. What SAP needs to do now is lay out a clear plan for bringing together these various components into a seamless Big Data platform and communicate the real-world benefits to customers. From what I have seen, SAP is well on its way to doing just that, but ultimately SAP customers will decide if the company has what it takes to survive in the Big Data market.

You can access the complete Wikibon Big Data Market Size and Vendor Revenues report here. Like all Wikibon research, access is free.

(Check out Jim Hagemann Snabe talking with SiliconANGLE’s John Furrier inside theCUBE about HANA and real-time analytics at last year’s SAPPHIRE.)