Holistic and Trajectory Approaches for Distributed Non-preemptive FP/DP^* Scheduling

Abstract

In this paper, we are interested in real-time flows requiring quantitative and deterministic Quality of Service (QoS) guarantees. We focus more particularly on two QoS parameters: the worst case end-to-end response time and jitter. We consider a non-preemptive scheduling of flows, called fp/dp $^{\rm {\sc *}}$, combining fixed priority and dynamic priority established on the first node visited in the network. Examples of such a scheduling are fp/fifo $^{\rm {\sc *}}$ and fp/edf $^{\rm {\sc *}}$. With any flow is associated a fixed priority denoting the importance of the flow from the user point of view. The arbritation between packets having the same fixed priority is done according to their dynamic priority. A classical approach used to compute the worst case end-to-end response time is the holistic one. We show that this approach leads to pessimistic upper bounds and propose the trajectory approach to improve the accuracy of the results. Indeed, the trajectory approach accounts for worst case scenarios experienced by a flow along its trajectory. It then eliminates scenarios that cannot occur.