A framework for timing analysis of event-driven applications

Abstract

Event-driven applications, particularly those based on the publish–subscribe communication model are widely adopted to build responsive and decoupled applications in domains such as robotics, the Internet of Things (IoT), and real-time control. While these architectures offer flexibility and scalability, they also pose significant challenges in meeting real-time timing requirements. These challenges often stem from factors such as long-running event-handler executions, misconfigured parameters, and execution orderings of event-handlers.

In this thesis, we address the problem: Given code based on the publish–subscribe communication model, will it consistently deliver messages on time?. To answer this, we propose a framework for analyzing the timing behaviour of such systems. Our approach involves constructing a Timed Automata Model from the source code, capturing both timing and behavioural semantics. The constructed Timed Automata Model is then verified using a model checker to determine if the system could ever end up in a situation where messages are not delivered on time.

If such a situation arises, the framework provides feedback such as recommending adjustments to configuration parameters, reordering event-handler execution, or relaxing overly strict timing requirements. Through case studies on several real-world ROS packages, a widely adopted publish–subscribe system, we demonstrate the practical utility of our approach.