Event-driven debugging for distributed smart applications

Smart applications are often written in an event-driven way, because the modularity of event-driven architectures aligns with the distributed nature of smart applications. Once event-driven applications get large enough, however, debugging them can become challenging. This is due to a lack of adequate debugging tools for event-driven smart applications. A key reason for this lack of debugging tools is a lack of a standardized model for debugging event-driven applications. Because no such model exists, runtimes do not provide many debugging functionalities specifically for event-driven applications. Additionally, event-driven applications are hard to debug because of the large state-space of possible executions that event-driven programs have. Furthermore, event-driven smart applications are often time-sensitive and also consist of multiple devices. To make it easier to debug distributed event-driven smart applications, we propose three research objectives. The first objective is to develop a debugging model for event-driven smart applications. The second objective consists of developing new debugging abstractions for debugging event-driven smart applications. The third objective is to invent novel state-space reduction techniques to reduce the number of explorable event sequences in the debugger. These three research objectives combined will result in a debugger capable of exploring the full state-space of possible events, making it much easier to find and reproduce bugs.

Keywords:Smart devices, Debugging, Internet of Things

Disciplines:Distributed systems, Programming languages and technologies, Embedded and real-time systems, Ubiquitous computing, Coding tools and techniques, testing and debugging