Diploma Thesis, TU-Vienna, 2014

The Peer-Based Programming Model (short: Peer Model) is an approach for designing and implementing distributed systems, developed by the Space-Based Computing group at the Vienna University of Technology. Its structure is inspired by asynchronous message queues, tuple space communication, staged event-driven architectures and datadriven workflows.

 

The Peer Model approach is relatively new and no tools that aid in the monitoring and debugging process exist yet. Developers need to resort to manual methods in order to gain insight into the workings of a running Peer Model based system, which is cumbersome and time consuming. This Master's thesis fills the gap by developing a visual monitoring software for the Peer Model approach, whose purpose is to help developers analyse and debug Peer Models.

 

A requirements analysis is performed to explore the needs and problems developers currently face when working with the peer model by conducting personal interviews with them. The analysis is strongly linked to the identification of what data is required to enable proper monitoring and debugging, as well as the development of suitable, standardised data formats. Suitable visualisations can facilitate the understanding and increase the productivity of developers better than textual representations. Fortunately, the Peer Model is a good fit for representing it visually. A graphical notation for representing Peer Models has been established prior to this thesis, but it only captures the static structure of models. Consequently, illustrations for the dynamic events occurring within a running Peer Model system are developed and implemented in the monitoring tool.

 

The visual peer model notation resembles nested graph structures. Automatic layouting of these structures is a crucial part of this Master's thesis as it spares developers from having to manually arrange elements in order to obtain an informative and visually pleasing representation. Thus, a methodology for layouting Peer Models is developed. Afterwards, the implementation of the monitoring tool is discussed and finally, the developed tool is critically evaluated and potential areas for future works are presented.