SAP Unveils HANA for In-Memory BI

May 24, 2011 Alex Woodie

SAP customers who need real-time analysis of their transactional data may want to take a look at HANA, the new in-memory business intelligence (BI) software officially announced at the ERP giant’s annual SAPPHIRE user conference last week. The offering, which includes software that’s pre-configured to run on hardware from partners like IBM, is said to provide a big shot in the arm for production reporting, with deployment in as few as six weeks.

In-memory databases have become popular over the last few years for analytic and reporting workloads due to their simpler implementations, lighter hardware requirements, and cost savings compared to traditional data warehousing setups. For starters, because the data to be analyzed resides in memory and not on a disk, queries can run much faster. Also, most in-memory databases are designed to run on X64 servers, which are cheaper than the big RISC-based servers that have traditionally been used to power big OLAP workloads.

SAP’s In-Memory Appliance, which goes by the colloquial HANA (for High-Performance Analytic Appliance), effectively serves as a “side mart” of data that’s separate from production workloads, with accessibility down to the line-item level, but perhaps only a minute (or less) behind the production system. This gives users the capability to run very detailed queries without impacting their ERP or CRM system.

Data stored by HANA can be accessed in several ways, including the pre-configured reports provided by SAP, which includes 15 operational reports across four areas of production, including sales, financial, shipping, and procurement. HANA-based data can also be queried by SAP’s BusinessObjects reporting solutions, or even through Microsoft Excel or other products for SQL or multi-dimensional expression (MDX) access. SAP envisions customers building their own planning, forecasting, operational performance, and simulation solutions on top of HANA.

HANA is pre-configured to automatically input the latest changes to customers’ SAP-based ERP and CRM systems, but users can feed data from outside sources as well, using BusinessObjects Data Services. As SAP says in its HANA brochure: “Organizations can instantly explore and analyze all of their transactional and analytical data in real time from virtually any data source.”

IBM i-based implementations of SAP are considered a native data source for HANA. An SAP spokesperson confirmed that HANA offers support for DB2/400, which is the database used by 1,600 SAP customers who run their enterprise software on IBM i-based servers. By the way, that installed base number is up from 1,100 five years ago, which shows SAP-on-IBM i is gaining momentum.

Apparently, however, HANA may not offer quick native support for SAP customers who run their systems on the Oracle database, which reportedly accounts for 60 percent of SAP’s massive installed base of 109,000 customers.

According to a report last week in InformationWeek, there is some confusion whether SAP’s Oracle-based customers will be allowed to access the database in real time. Without the licensing snafu cleared up, the only access point would be via batch loads, which sorta defeats the purpose of a real-time analytical system. Oracle, of course, is a competitor to SAP in the ERP and BI spaces–and Oracle offers its own in-memory database, as well as its Exadata appliances, which is seen as a direct competitor to HANA.

HANA has been in limited release since last year, with general availability expected in July. One customer that’s been testing HANA since December is Colgate-Palmolive. The company’s CIO, Tom Greene, says Colgate-Palmolive had difficulty analyzing SKU-level sales by retail location.

“Prior to SAP HANA, we were unable to run full analytics in a reasonable timeframe,” Greene states in a press release. “With SAP HANA, we will be able to run analytics at a local level on specific brands and locations, and at the lowest level of detail in real time.”

IBM and SAP have worked to certify HANA to run on two Novell SuSE Linux-based System x servers, the x3690 X5 and the x3950 X5. The x3690 is a 2U server equipped with 256 GB of memory, two Intel Xeon processors, and the choice of pure solid state drives (SSDs) or a mix of SSDs and traditional spinning drives. The x3950 is a 4U server equipped with 256 GB or 512 GB of memory, two or four Xeon processors, and the choice of SSDs or hard drives.

For more information on HANA, see www12.sap.com/platform/in-memory-computing/in-memory-appliance/index.epx.