Task partitioning of incompletely specified real-time distributed processing systems

Real-time applications often demand a distributed processing system due to speed, processing power, time constraints etc. Optimizing the number of processors (for a dedicated system) is thus an important issue both from the economic and performance standpoints and can be carried out during the task allocation process. This aims at mapping the software modules of the real-time application onto the distributed processing system such that time constraints are met using minimum system resources (number of processors). Thus time specifications play a critical role in the allocation process. In this paper we present an extension of an existing task allocation scheme to allow the system designer to incompletely specify the time specifications. The original scheme required that time constraints be specified for all nodes. Reduced specifications correspond more closely to real design need, but can require a more exhaustive search for a near optimal solution. This work describes an efficient approach for this case.