This thesis conducts a rigorous comparative analysis of diverse tracing mechanisms in the Linux ecosystem, with a specific focus on their suitability for real-time systems. It investigates the fundamental principles of tracing methods such as ftrace, perf, eBPF, and LTTng, assessing their impact on system overhead, latency, and accuracy. Performance metrics encompassing context-switch overhead, event capture efficiency, and real-time responsiveness are used to evaluate their efficacy across various real-time workloads.
The study also delves into seamless integration strategies for these tracing mechanisms within real-time applications, addressing challenges related to interference with critical tasks and optimizing trace collection. The findings equip developers, system architects, and researchers with actionable insights for selecting appropriate tracing tools in diverse real-time scenarios.
In essence, this thesis offers a comprehensive understanding of tracing mechanisms' compatibility within real-time Linux systems. It aims to enhance the development of efficient tracing solutions for demanding real-time applications, ultimately contributing to improved system performance and stability.