Analyzing Distributed Remote Process Execution Using Queuing Model

Abstract

The mechanism of executing processes at remote nodes in a distributed system enables the optimal utilization of computing resources. The main aim of distributed remote processing is to utilize idle resources of a node in a distributed system. However, the arrival of remote processes in an idle node is unpredictable. In addition, the execution of remote processes along with existing local processes in a node is required to provide enhanced performance without thrashing due to enhanced process-load. In this paper, an algorithmic model of distributed remote processing is constructed. The properties of the distributed remote processing are analyzed using different queuing models. The stability criteria and the performance metrics are derived following Markovian and Discrete-Time Markovian Arrival Process models. The experimental results illustrate that proposed algorithmic model can improve the resource utilization of an idle node through distributed remote processing.