This project is exploring new communication mechanisms that go beyond the deadline guarantees of previous work in real-time and non-real-time message streams, using models of real-time communication that can represent tolerance to deadline misses. For real-time message streams, these models suggest new techniques for providing service guarantees on the frequency or clustering of missed deadlines. for non-real-time message streams, the models will be used to develop support mechanisms that minimize average delivery delay subject to the constraints of the real-time streams. Simulation and analytic methods are being used to evaluate competing techniques.
