Ultrafast Analytics

Feature Article | November 7, 2012 by Torben Hügens and Matthias Merz from Camelot ITLab

Illustration: grasundsterne

The in-memory computing technology SAP HANA is designed for high performance: data is stored in columns in the main memory, it has sophisticated data compression and retrieval techniques, and database tables are partitioned. When running on an optimized hardware platform, SAP HANA offers efficient parallel processing of database queries. Compared to conventional database management systems, performance gains of a factor of 10 to 100 can be achieved; in certain scenarios gains of a factor of 1,000, 10,000, or more, are possible. There are two SAP HANA scenarios – integrated and side by side.

2 SAP HANA scenarios: Integrated and side by side

The integrated scenario covers products that use SAP HANA as the primary database, such as SAP Business One Analytics, powered by SAP HANA, and SAP NetWeaver BW, powered by SAP HANA. More “powered by SAP HANA” solutions have been announced for the fourth quarter of 2012, including SAP CRM, SAP ERP, and SAP SCM.

In side-by-side scenarios, an SAP HANA database is connected to an existing database. Selected data is replicated so that separate applications can be built on it, data can be analyzed in real time (SAP HANA as a data mart), or to accelerate certain processes (SAP HANA as an accelerator).

In this article we present the most important tools in the SAP BusinessObjects BI platform for analyzing data in both scenarios, and explain how they interact with SAP HANA. But first we need to explain the differences between BW powered by SAP HANA and SAP HANA as a data mart.

BW powered by HANA

In this scenario, SAP NetWeaver Business Warehouse (BW) uses the SAP HANA appliance software as the primary database. Having the data stored in columns in the main memory means that measures, or columns, can be read much faster, and totals and averages can be calculated quickly – even for vast numbers of data records. InfoProviders designed specifically for SAP HANA, such as DataStore objects and InfoCubes optimized for SAP HANA, further accelerate the loading and analysis of data in BW, since complex and performance-intensive processes, such as activating DSO requests, can be done in the SAP HANA appliance software itself.

SAP HANA as a data mart

In this deployment scenario, the SAP HANA appliance software is used alongside an existing database. Operational data from SAP or non-SAP systems can be replicated to the SAP HANA database using the SAP LT Replication Server or SAP BusinessObjects Data Services. Whereas SAP BusinessObjects Data Services is used to set up complex processes to extract, transform, and load data, the SAP LT Replication Server brings about a trigger-based replication of all relevant tables. When data is inserted or updated in the ERP system, it is automatically transmitted to the SAP HANA database so that it is available for almost real-time reporting. Data in the SAP HANA appliance software is accessed using information models such as attribute, analytic, and calculation views, which can be created in the SAP HANA studio.

The SAP BusinessObjects BI platform can be used as the main reporting platform in both scenarios. Various tools can be used to meet almost all reporting requirements. See chart below for more details.

Chart from Camelot ITLab

Which reporting tools should you use

In the BW powered by SAP HANA scenario, SAP NetWeaver BW is still the main data warehouse, and data is retrieved using the query design tool BEx Query Designer. With SAP BusinessObjects BI platform 4.0 SP4, all client tools can access BEx queries using the BI Consumer Services (BICS) interface. There are some features that the interface does not support, such as certain aggregations. An OLAP connection in the central management console of the SAP BusinessObjects BI platform is required to access BEx queries. With this connection in place, the BEx queries can be used in all client tools. Exceptions are SAP Visual Intelligence and SAP BusinessObjects Explorer. SAP Visual Intelligence can currently use only SAP HANA, relational databases (via JDBC), MS Excel spreadsheets, and relational UNX universes as data sources. SAP BusinessObjects Explorer cannot yet access SAP NetWeaver BW InfoProviders directly if SAP HANA is the primary database.

In the SAP HANA as a data mart scenario, SAP NetWeaver BW is not needed, as the SAP HANA appliance software is the primary database. Data is stored and modeled in the SAP HANA database. The analytic and calculation views described above are used to deliver the data to the client tools. These views can be accessed by creating an OLAP connection in the central management console (for SAP BusinessObjects Analysis, edition for Microsoft Office and OLAP, and SAP Crystal Reports for Enterprise). The client tools can also access the data in the SAP HANA appliance software using the OLAP connection. Depending on which client tool is used, different technologies are needed to access the data. In some cases, a relational universe (.UNX) has to be used (for SAP BusinessObjects Web Intelligence and dashboards).

Whichever SAP HANA scenario is deployed, users can choose from an array of tools for reporting. When SAP HANA is used as a data mart, SAP Visual Intelligence is an additional tool with which users can create attractive visualizations. They can also produce their own reports to obtain more business insights.

All in all, the combination of SAP HANA and the reporting capabilities of the SAP BusinessObjects Business Intelligence platform meets all analytics requirements without the performance bottlenecks associated with large volumes of data.

Tags: ,

Leave a Reply