Over a decade ago, Leonard Kleinrock developed a model for selecting best–performance scheduling algorithms for computer systems. The M/G/I queuing system used by Kleinrock requires that an estimate of the service–time probability distribution function (p.d.f.) be made in order to derive results. While methods of fitting service–time data to p.d.f.s are available, this paper argues that the use of a generalized beta distribution of the second type (GB2) in concert with the Kleinrock model reduces the need to estimate alternative models and avoids the consequences of arbitrarily imposing invalid assumptions on the data.