SAP和SAS为获取快速大数据分析形成联盟

藤菜

原文链接：http://www.forbes.com/sites/tomg ... big-data-analytics/
SAP and SAS , often regarded as competitors, have agreed on a partnership to run SAS analytics on SAP’s very fast in-memory HANA platform. The announcement had many people in the business surprised and puzzled at why these two firms had decided to cooperate now. After all, SAP undertook the purchase of a predictive analytics company, KXEN, just last month. KXEN competes with SAS and with IBM IBM +0.28%’s SPSS.  SAP also owns BusinessObjects, a business intelligence and reporting system.  And only a year ago, a SAP executive said, SAS was publicly trashing HANA.

“I think it is a win-win for the high performance analytics space,” said James Taylor, CEO of Decision Management Solutions, a consultancy. “In-database analytics is a powerful approach that helps scale and industrialize advanced analytics – especially predictive analytics. SAS is extremely well established in advanced analytics and SAP HANA is making a strong play to be a very high performance data environment. It makes sense for both companies to work together to deliver high performance in-database analytics.”

Ken Tsai, vice president of HANA marketing, cast the agreement as a way to offer customers choice in analytics on the HANA platform, part of SAP’s strategy to offer a portfolio of tools on its engine. SAS has been dominant in mathematical calculations — the sort that appeal to, and often require, analysts with Ph.D.’s, while SAP aims its analytical tools at business users, he said.
Russ Cobb, vice president of alliances and product marketing at SAS, also explained the alliance as a response to customers. When SAP first launched HANA, it wasn’t clear what the use cases would be, said Cobb, and some thought HANA was a competitor to SAS.

“As we understood HANA and they understood what we were doing with our in-memory work, it showed they were complementary. We had joint customers who were asking us when we were going to get our act together and work with each other — customers with big SAP implementations who were starting to do HANA and had SAS as well.”

Part of the impetus on the part of SAS is that it needs some high speed engine because it runs on some old architecture that can be pretty slow, said Peter Sollimo, vice president for sales at Fuzzy Logic, a fast competitor. One of the problems with SAS, according to Sollimo and others, is that it often moves data from storage to a SAS processing engine, which takes time, especially with large amounts of data. In addition, many developers who are comfortable with SAS use it for everything, including data preparation and data aggregation where it is not especially efficient.  Doing data prep in Teradata TDC +3.19% before sending it to SAS for analytics can improve performance 10-20x, Sollimo said. A Teradata partner, SAS can run in Teradata’s database and Teradata has its own in-memory analytical applicatin.

SAP’s  Tsai said the company has a longer term strategy to create industry solutions with SAS.

“SAS has portfolios of apps in multiple industries. Part of the agreement is to take this opportunity and identify which industries make sense to offer to the companies,” he added. “This is a pilot. We will select a few customers and we will make sure that we validate the requirements and commercial opportunities. Our goal is by early next year we will have solutions built on the HANA platform using SAS analytics and apps.”

The SAS agreement is part of SAP’s opening up to other vendors, added Tsai.

“It used to be only SAP-built and open source. Now we are extending the platform strategy to one of the best of breed solution providers in the market,” Tsai explained. “This doesn’t mean we don’t have ambitions to present our own predictive analytics.”

“Clearly they will continue to compete,” added Taylor “SAP won’t stop pushing BusinessObjects or its Predictive Analysis tool or its new KXEN acquisition. Nor will it stop making sure that these products work well with SAP HANA. SAS meanwhile will continue to develop its high performance analytics strategy and partner with other vendors for in-database analytics. Companies face an increasingly complex future where multiple analytic tools and multiple data platforms are the norm so this kind of co-opetition is key.”