Resource budgeting as a tool for reduced development cost for embedded real-time computer systems
Abstract: Wouldn’t it be great if there were a systematic method for derivation of non functional con-straints available at design time that made it possible to verify design and make implementa-tion a much clearer task? This kind of methods are needed since systems of increasing com-plexity has to be developed, and the cost for failing has proven to bee too high. The problemis how to derive the design time constraints into implementation time constraints, maintainingthe traceability for the individual constraints, and early on get indications whether a project isabout to fail or not.A method for implementation time constraint derivation has been developed and is presentedin this thesis. Along with the basic method, several extensions are proposed. Evaluationsof the practical usefulness of the method and the method’s scalability have been done. Toprove the method’s importance in real development projects, a method for evaluation of theusability of this kind of methods has also been developed. The evaluation of the practicalityshows that it is possible to find close to optimal solutions (within percent) in short time (withinminutes). The evaluation of the scalability shows that the run time for finding implementablesolutions scales polynomial with the size of the task graph. The evaluation of the usabilityshows that using the proposed method always leads to lower development cost than using anad hoc method, in the case that the implementation is about to fail.
This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.