What is it about?
Based on execution traces we extract a model of the underlying behaviour for understanding and diagnosing problems. For repeated behaviour sequences we can automatically identify outliers. Either focusing on the order in which events occur or the timing between them
Featured Image
Photo by Erik Mclean on Unsplash
Why is it important?
In complex situations with many parallel action sequences dependencies between the actions may cause dramatic performance problems. Analyzing these problems is very hard without ways to visualize the behaviour. Automatic detection of outliers also helps. We have used this in problem analysis where domain experts spent weeks without detecting a root cause, whereas with our method we were able to isolate the problem is a couple of hours.
Perspectives
Relevant for any system that has parallel complex behavior
Jos Hegge
TNO
Read the Original
This page is a summary of: Platform Performance Suite (PPS): A Framework for Performance Analysis & Diagnosis of Complex Cyber-physical Systems, May 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3676151.3719356.
You can read the full text:
Contributors
The following have contributed to this page
