Workload bounds in fluid models with priorities

In this paper we establish upper and lower bounds on the steady-state per-c lass workload distributions in a single-server queue with multiple priority classes. Motivated by communication network applications, the model has co nstant processing rate and general input processes with stationary incremen ts. The bounds involve corresponding quantities in related models with the first-come first-served discipline. We apply the bounds to support a new no tion of effective bandwidths for multi-class systems with priorities. We al so apply the lower bound to obtain sufficient conditions for the workload d istributions to have heavy tails. (C) 2000 Elsevier Science B.V. All rights reserved.