Partitioned Scheduling of P-FRP in Symmetric Homogenous Multiprocessors

Functional Reactive Programming (FRP) is a declarative approach to modeling and building reactive systems. Priority-based FRP (P-FRP) has recently been introduced as a FRP formalism that guarantees real-time response. P-FRP guarantees that when a higher priority task is released, the system will immediately preempt any executing lower-priority tasks. To maintain guarantees of state-less execution offered by the functional programming model, P-FRP implements a transactional nature of execution. Each higher priority event in P-FRP can abort a lower priority task forcing it to restart. Existing work on partitioning tasks in multi-processor systems have been focused on the classical preemptive model of execution1. However, due do its transactional nature, the schedulability tests used in the partitioning algorithms for the preemptive model, cannot be applied 'as is' to the P-FRP execution model. While multiprocessor response time analysis of P-FRP has been done in previous work, partitioning schemes for tasks in multi-processor systems have not been presented yet. In this paper, we present an exact schedulability test for P-FRP and use it in two existing first-fit partitioning schemes. We also introduce a new first-fit partitioning scheme based on the processing time of tasks, which yields better results than the other two schemes. We also show that the number of processors required to schedule tasks in P-FRP are more than or equal to the number of processors required to schedule the same in the preemptive model.