Functional model of a software system with random time horizon

Abstract

Virtualization technologies are being actively used to design infrastructure of cloud computing systems. In this case applications can be duplicated and hosted in different virtual machines on different physical nodes. That defines various performance of applications which causes the problem of managing performance of the entire heterogeneous system. There are different ways of solving this problem, including queuing theory methods. However research of the threshold discipline in scope of queuing theory is not complete because of difficulty of gathering precise analytic values and building of precise mathematic model of the system. Another feature of heterogeneous systems is the finite random time of system functioning which is defined by random endogenous and exogenous factors. This paper gives an overview on a functional model of the system with two heterogeneous devices with random functioning time and different service disciplines. In scope of simulation statistic experiments for different service disciplines at random time interval an average time needed to process a single request is measured. A comparison of service disciplines is conducted. Authors also provide a working software implementation of the heterogeneous system and experiments with use of service disciplines is performed.