Consider the model with nonpreemptive priorities presented in Sec. 17.8. Suppose there are two priority classes, with λ1 = 2 and λ2 = 3. In designing this queueing system, you are offered the choice between the following alternatives: (1) one fast server (μ = 6) and (2) two slow servers (μ = 3).
Compare these alternatives with the usual four mean measures of performance (W, L, Wq, Lq) for the individual priority classes (W1, W2, L1, L2, and so forth). Which alternative is preferred if your primary concern is expected waiting time in the system for priority class 1 (W1)? Which is preferred if your primary concern is expected waiting time in the queue for priority class 1?