This course stresses the application of queueing network models as tools for answering important questions of cost and performance that are routinely confronted by professional computer system performance analysts. The models considered can be used throughout the life of a computer system: during design and implementation, during sizing and acquistion, and during evolution of configuration and workload. An operational approach to queueing network modelling and case studies are well-integrated into the course. For models of systems having finite-state space, we also consider Markov chain performance models. An preliminary outline of topics is attached.