TRACE User manual

This documentation describes on a high level how the TRACE tool can be used. Although the techniques are tailored to performance engineering of cyber-physical systems, the applicability of the TRACE tool is broader than that. For formal explanations of the concepts and analysis techniques, we refer to the several peer-reviewed academic publications in the references. We have created a number of example traces for the running example that are available here.

Contents:

• Introduction: an introduction to the TRACE methodology.
• Running example: description of a running example that is used to explain the various techniques. It contains a link to a zip file that has several example traces.
• Visualisation: an explanation of the TRACE visualisation features.
• Critical-path analysis: an explanation of the critical-path analysis supported by TRACE. This can be useful for finding performance bottlenecks.
• Distance analysis: an explanation of a pseudo-metric on traces. The technique can be used to find and highlight differences between similar traces. This can be useful during model calibration, i.e., when comparing a model trace with a system trace.
• Resource-usage analysis: an explanation of the resource-usage analysis supported by TRACE.
• Little's-Law analysis: an explanation of the TRACE version of Little's Law that relates throughput, latency and work-in-progress.
• Behavioral analysis: explanation of a technique that can be used to find anomalies in repetitive behavior within an execution trace.
• Runtime verification: explanation of the support for runtime verification.
• File format: explanation of the TRACE input file format.
• References: references for a more detailed and formal explanation of the TRACE techniques.