Speed up SAP Netweaver Business Intelligence queries using IBM PureData System for Analytics (Part 2)

Imagine that you’re the CIO of a big retailer with both an online and store presence. Christmas season is coming, and last year around this time, your critical business processes almost broke because of the heavy demand for deep analytics and frequent reporting during the high season. What would you do?

In the first post of this series, I described an alternative reporting and analytics solution to SAP Netweaver Business Intelligence, based on IBM PureData System for Analytics, powered by Netezza technology. In order to provide business users with the analytical performance they need to keep up with the growing demand for more data, faster reports, deeper analytics and less cost.

The challenge

A large retail client in Germany had exactly the problem I introduced in the beginning of this post. Their SAP Netweaver Business Intelligence system is essential for certain business critical processes such as month end closure, inventory and demand planning. Under normal conditions, the reporting and analytical workload can be processed in parallel to the more transaction-oriented processes.

But every Monday just before and just after holiday seasons such as Christmas, the system came very close to its maximum capacity—sometimes to capacity overload.

The evaluation process

What would be the best way to tackle the challenge? First, the client needed to evaluate the technical fesibility, the time it would take for a successful implementation and the associated costs.After the evaluation of the different possible approaches, such as upgrading SAP Netweaver Business Intelligence to the latest release, implementing SAP HANA or changing the underlying database out with another, the client decided to implement a sidecar solution to offload the most critical reports and analytics to reduce the workload on SAP Netweaver Business Intelligence.

The solution

The idea was to introduce a purpose built, easy-to-use, high-speed analytics data warehouse appliance that grows with the business requirements: PureData System for Analytics.

Implementation of such a sidecar solution is a best practice approach also recommended by SAP itself with their Sybase IQ columnar database.

After a one week assessment, the top three SAP Netweaver Business Intelligence reports had been identified and run by thousands of users several times every day. Removing this purely analytical workload from the system would guarantee a smooth season next Christmas.

The data (several terabytes) had been integrated with the help of a business partner using the customer’s incumbent extract, transform and load (ETL) platform, which could be connected to SAP and PureData System for Analytics.

For reporting, the client introduced IBM Cognos along with PureData System for Analytics, gaining the maximum out of the new analytics infrastructure.

The result

The most important fact is that our client survived Christmas season (and Easter as well).

Their SAP Netweaver Business Intelligence system can still serve its purpose, is running smoothly and has been very stable since then. Only the reporting and analytics run now on the sidecar PureData System for Analytics. The response time for typical queries is mostly under two seconds.

Because of the highly flexible implementation of data model and granularity within PureData System for Analytics, the client was even able to increase the frequency of some reports from monthly to weekly updates, which enabled the business users to do more with less effort in a shorter amount of time.

The retailer started the implementation in April 2013 and finished the project in September 2013, on time and on budget.

What do you think of the implementation? Are you facing similar challenges? Let’s connect and follow me on Twitter.