One of the main lessons of queueing theory (Chaps. 17 and 18) is that the amount of variability in the service times and interarrival times has a substantial impact on the measures of performance of the queueing system. Significantly decreasing variability helps considerably. This phenomenon is well illustrated by the M/G/1 queueing model presented at the beginning of Sec. 17.7. For this model, the four fundamental measures of performance (L, Lq, W, and Wq) are expressed directly in terms of the variance of service times (σ2) so we can see immediately what the impact of decreasing σ2 would be. Consider an M/G/1 queueing system with mean arrival rate λ = 0.8 and mean service rate μ = 1, so the utilization factor is 
