Worst-Case Measurement-Based Statistical Tool
Pavel G. Zaykov, Jan Kubalcik · 2019
In this paper, we address the problem of computing the worst-case timing bound for avionics applications. The worst-case bounds are defined by the application Worst-Case Execution Time (WCET). The avionic applications are executed on a Commercial-Of-The-Shelf (COTS)multi-core processor from the embedded domain. Based on the identified gap in the existing state-of-the-art approaches analysis, we propose a WOrst-case Measurement-Based stAtistical Tool (WOMBAT). The proposed WOMBAT uses the Extreme Value Theory (EVT). WOMBAT is a stand-alone tool, and it relies on the user to correctly configure the experimental setup. As an input, WOM-BAT employs traces from application execution time traces and the probability of exceedance defined by the Design Assurance Level (DAL). As an output, WOMBAT estimates the application WCET. We experiment with execution time traces from real and synthetic applications, both executed on an embedded COTS multi-core processor. The purpose of the experiments is to demonstrate the applicability of the proposed statistical approach.