In our experience, SAP HANA in-memory technology leads to significantly better performance in a variety of usage scenarios. Nevertheless, most customers tend to use SAP HANA in one of two ways: either for operational analytics to support operational reporting, or in combination with SAP Business Warehouse. SAP expert Dr. Matthias Merz, explains how these scenarios work on a technical level.
1. SAP HANA as a standalone database
In the first scenario, SAP HANA functions as a standalone database to support operational reporting and runs in parallel with existing application systems. With the help of the SAP LT Replication Server (SLT) or SAP Business Objects Data Services (BODS), operational enterprise data can be replicated from SAP and non-SAP systems into the SAP HANA appliance. While BODS enable the creation of complex ETL (Extract, Transform, Load) processes for extracting, modifying, and loading data, the SLT allows a trigger-based replication of all the relevant tables. A specially designed SAP add-on transmits datasets simultaneously to the HANA database when insert or update operations are executed on tables in an ERP system, which means that the datasets are available for reporting in near real time.
Once the enterprise data to be analyzed is available in the HANA appliance, the next step is to enable operational reporting. Generally speaking, this scenario is based on the use of the BusinessObjects frontend tools, which access SAP HANA via special interfaces. Access to the data in the HANA appliance takes place via information models (attribute, analytic, and calculation views), which can be created using the SAP HANA studio.
2. SAP HANA as the primary database in a BW system
The second usage scenario involves using the HANA appliance as a primary database in a SAP NetWeaver BW system. Starting with SAP NetWeaver Release 7.3, SP5, you can perform a heterogeneous system copy and replace your existing database (such as Oracle, DB2, or MS SQL) with SAP HANA. In this scenario, the SAP NetWeaver application server and the SAP HANA database run on different hardware systems. All the data from the BW system is stored in the HANA appliance’s main memory. To protect the data from potential loss due to incidents such as power outages or hardware defects, conventional database techniques are integrated in the HAHA appliance, making it possible to create snapshots and write native database logs. This approach has already been used successfully in the context of SAP liveCache technology for many years.
The ideal scenario: SAP HANA in a BW system
The advantages of using SAP HANA as a database in a BW system are plentiful. As well as benefiting from high-speed access to data – thanks to in-memory technology – you no longer need to create aggregates or build indexes on InfoCubes. Data-loading processes are simpler and execution is significantly quicker. Thanks to the column-based data storage approach, you have rapid access to individual key figures and you can model aggregate functions such as “total figure” and “average figure”, which is particularly important if you are working with a large number of datasets. As a result, you do not need to use other hardware solutions to accelerate reporting. In fact, we anticipate that SAP HANA will replace SAP BW Accelerator completely.
Further performance gains can be achieved by shifting complex calculation operations to the database layer and adapting the application logic accordingly. SAP is realizing this approach through the SAP HANA-optimized InfoProviders that it has introduced in BW 7.3. Using an “in-memory- optimized InfoCube”, for example, leads to enormous speed gains when loading data, because the SAP HANA technology has made it possible to significantly simplify the data model for this type of InfoCube. At the same time, improvements have been made to the delta-set generation logic in the SAP HANA-optimized DataStore objects to make the data activation process much faster.
Applying such attributes as “powerful” and “best-in-class”, SAP is keen to emphasize that customers can lower their operating costs by deploying SAP HANA. Only time will tell whether this really is the case or not. After all, it is worth bearing in mind that deploying SAP HANA involves one-time migration costs, as well as new licensing and hardware costs. And, in many cases, employees who are responsible for the BW system require training.