I recently taught a roomful of 15 students how to use SAP BusinessObjects Web Intelligence BI 2025. One of the exercises focused on Data Mode, Web Intelligence’s new self-service data preparation and blending feature. Simultaneously, 15 students and an instructor combined two data queries and clicked the final “OK” button. A few seconds later, every session crashed.
So, we repeated the exercise.
BOOM!
The server crashed again.
The third time, my students and I were able to quickly save our Web Intelligence documents before the server crashed.
This behavior was something we had not seen the previous month with a class of only five students. And it reminded me of a few things.
First, it’s unlikely that fifteen users will perform the same sequence of operations at the same time outside of a training room. However, it is likely that fifteen (or more) users will be using Data Mode at the same time.
Second, just because something is unlikely doesn’t mean it shouldn’t be tested. The incident reminded me of an Airbus A380 documentary my son, an aspiring teenaged pilot, watched. During testing, engineers deliberately flushed all of the aircraft’s toilets simultaneously to verify that the plumbing system could withstand an extreme but conceivable workload. You can watch the clip on YouTube.
Third, the fact that a classroom-sized workload could consistently crash the system highlights a larger issue: SAP’s BusinessObjects sizing guidance has not kept pace with the platform. The current sizing guide predates the introduction of Data Mode and provides little insight into how concurrent Data Mode usage affects Tomcat, the Web Intelligence Processing Server (WIPS), memory consumption, CPU utilization, or overall platform scalability. A quick glance in the Central Management Console confirms that SAP did not introduce a new Adaptive Processing Server service specifically for Data Mode, but publicly available documentation offers few clues regarding where the workload is actually performed or how much additional capacity it requires.
Perhaps most importantly, customers have no published guidance for answering a simple question:
How many concurrent users can perform Data Mode operations before performance begins to degrade?
Without updated sizing benchmarks, administrators are left to answer that question through trial and error. SAP should publish updated capacity-planning and load-testing guidance that reflects modern Web Intelligence workloads, including Data Mode. Customers should not have to discover scalability limits in the middle of a training class.
Let’s hope that SAP will revise the sizing guide in time for the BI 2027 product launch, which is expected in Q1 2027.
How are you sizing and capacity planning for Web Intelligence Data Mode in your environment?
