What to do when SAP BW is reaching its limit?

What options do you have when you start running out of your BW system storage space? Let’s see how one of our customers faced this challenge.

Duerr, a global mechanical and plant engineering company from Germany, started to reach the threshold of its BW system. This was accompanied by performance bottlenecks, occurring naturally when HANA starts to reach its allocation limit. In such cases users typically experience performance problems such as slow response times and volatility in query runtimes. This is caused by higher table unloads which is how HANA responds when it needs more space for its calculation processes.

In such cases, extending the size of your database or deploying archiving strategies often seem like the only solutions. Of course, neither of these solutions is fast or easy. So for our customer we found a third option that helped them save 30% of storage fast and without costly investments.

A thorough system check points to the best solution

To help Duerr, our consultants started with an implementation of BW Fitness Test, a tool providing an extensive overview of BW systems. The analysis is a great starting point especially for data management and performance topics. Considering the goal of the analysis, we focused on identifying “frozen data”. Frozen data is technical data which quickly lose its value. It can therefore be deleted and free up your memory space.

Datavard FitnessTest monitors data usage and calculates its distribution in a productive system, while checking other performance related metrics. The test often sheds a new light on your system and points to unexpected correlations. After analyzing the test results, our consultants prepare and present recommendations for further actions aimed at improving the system’s performance.

Alternative strategy revealed

In this case, the analysis showed generous 300 GB of frozen data. This meant a significant housekeeping potential amounting to almost one third of the system. Thanks to proper housekeeping, the memory extension can be postponed by 3-4 years.

Other measures, requiring relatively little effort but bringing considerable benefits in terms of system’s size and performance (,,quick wins“) were also recommended. Datavard consultants also presented a detailed temperature analysis of all reporting data in the customer’s system. Such analysis serves as a starting point for further data offloading projects but can also be used for efficient performance tuning.

Find out more and read the whole story here.