A sizing question runs through the model in four steps. First, state the demand roughly — for example, a service that handles on the order of a few hundred requests in a peak minute. Second, state how bursty that demand is, in coarse terms such as mostly steady with occasional spikes. Third, let the model produce an operating range for the resource, say a band of capacity rather than a single figure. Fourth, read that band as a decision input: the low end is what handles ordinary load, the high end is what absorbs the spikes. The value of the walkthrough is that no step required a precise measurement, and the answer still supports a real decision about how much to provision.
The Jev Model: An Intuitive Overview