Histogram - Based Performance Analysis for Computer

A conventional analytic model used for evaluating the performance of computer and communication systems accepts single values as model inputs and computes a single value for each performance measure of interest. However uncertainties regarding parameter values exist in diierent situations such as during early stages of system design. Although the clients in a system are statistically identical factors such as the time of the day and the current size of the data les can introduce variabilities in service demands for the server devices. Existence of uncertainties or variabilities in service demands makes the use of a single mean value for each model parameter inappropriate causing the conventional modelling approach to become ineeective. This paper proposes to use histograms for characterizing one or more model parameters that are associated with such uncertainty or variability and demonstrates its application with separable queueing network models. A histogram consists of a number of intervals (ranges) with a probability of occurrence associated with each interval. The adaptation of the existing solution techniques for both open and closed queueing networks to histogram parameters is presented in the paper. One or more input parameters for the model can be speciied as a histogram and a histogram is generated for the model output. Such a histogram output is also appropriate when percentile values for performance measures are required in addition to the mean value. Examples of the application of these techniques to diierent problem domains and the computational complexity of the technique are discussed. The cost associated with running the algorithm is observed to be reasonable as long as the number of devices whose demands are speciied as histograms is not too high. Although the algorithm produces accurate results in many situations inaccuracies have been observed for certain systems. A technique called interval splitting that can be used for controlling such inaccuracies is described. 1 (per sec) 2 (per sec) (per sec) RT (sec) RT 1 (sec) RT 2 (sec) RT (sec) (period 1) (period 2) (aggregated) (aggregated) (period 1) (period 2) (overall)