Tolga Tezcan. Analysis of the flexibility of scheduling rules using augmented fluid models

Abstract. In this talk we focus on the flexibility of scheduling policies. More specifically, we would like to answer the following question; if we fix a policy and introduce more flexibility into the system does it always perform better? We show that commonly used scheduling policies such as static priority policies, e.g. cmu policy, longest queue first and FIFO are not flexible. In fact, we show that adding more flexibility to a system (without decreasing the capacity) working under these policies can make the system become unstable. The analysis of these systems with traditional fluid models is not possible. We introduce the augmented fluid models and prove stability or instability of these systems using these models.