SAP's High-Performance Analytic Appliance , or HANA, which started shipping on Wednesday, should give its customers a much faster, more flexible way to analyze large volumes of data in real-time, analysts said.
Much of the performance gains come from HANA's use of an in-memory computing technology which allows data to be processed in a system's RAM as opposed to reading it off I/O disks. The in-memory approach enables much faster data processing and should allow companies to run far more sophisticated data analytics applications compared to conventional relational databases.
"Any enterprise that is a user of large data sets for data analytics," will find HANA interesting said Frank Scavo, manging partner of the IT consulting firm Strativa. "HANA represents a large improvement not only in processing speed but in flexibility and cost of data analytics."
HANA's ability to let large data sets be manipulated in memory will enable enterprises to run more ad hoc queries against business data, and reduce the need for pre-define cubes and queries, he said.
Companies will be able to take a large dataset and decide what questions they want to run against it, in real time. "It adds a lot of flexibility to the whole process of data analysis," Scavo said.
In contrast, with most conventional data warehouses where data is stored in static tables "you have to predict the sort of questions you will ask about the data so you can architect it accordingly," Scavo said.
SAP said that during pilot tests with a consumer goods company, HANA's in-memory computing engine enabled the company to run arbitrary, complex queries on over 450 billion records in a matter of seconds, compared with the hours it would have taken previously.
Such capabilities are becoming increasingly useful, especially in industries such as the retail sector, health care, scientific analysis and financial services, Scavo said. "There has been an explosion in the amount of structured and unstructured data that organizations have been able to collect," and are looking to mine, he said.
SAP is not the first one with such technology, said David Menninger, an analyst with Ventana Research. Vendor interest in in-memory technologies has been growing over the past few years, at least in part because of the widespread availability of relatively low cost 64-bit hardware and operating systems, as well as low cost storage , he said.
Other vendors with in-memory technologies include QlikView, Tibco, Tableau, TM1 which IBM acquired through its purchase of Cognos, and Quantrix, he said.
HANA's difference lies in its scalability and the fact that it is being offered by a vendor that also owns a lot of the ERP, CRM and other data that companies are looking to analyze, Menninger said. "SAP has an advantage the others don't have," Menninger said. "They have access to the data from the sources systems."
The ability to tie HANA tightly into existing SAP -powered applications and transaction data is something that will make the technology appealing to SAP users, he said. "They control a lot of the applications and the data associated with the applications."
SAP said HANA is designed to work in non-SAP environments, and with data from any source. But the technology is likely to resonate mostly within SAP's own customer base said Merv Adrian, an analyst with Gartner.
Jaikumar Vijayan covers data security and privacy issues, financial services security and e-voting for Computerworld. Follow Jaikumar on Twitter at @jaivijayan or subscribe to Jaikumar's RSS feed . His e-mail address is email@example.com .
SAP has an uphill fight to sell HANA outside its base, Adrian said. It's a crowded market, they are entering it late, and not perceived as playing except with existing customers. They will need to build references, a sales force, and a partner ecosystem in a new field -- regardless of the quality of the product on first ship. It will take a while, he predicted. Keywords:
Read more about bi and analytics in Computerworld's BI and Analytics Topic Center.
This story, "SAP's HANA will speed real-time data analytics" was originally published by Computerworld.