Performance modeling of cloud e-marketplace using dynamic control model

Abstract

One major issue that has been of a great challenge in cloud E-marketplaces is the effective management of resources. To satisfy consumers, some E-market providers have adopted the resource over provisioning model. But literature reveals that resource over provisioning could lead to a largely sub optimal utilization of the hosting environment thereby leading to unnecessary wasteful costs. Also, with server consolidation, about $45 billion was spent on server management and administrative costs in 2012 alone. Therefore, effective resource management of these Server Machines in terms of cost minimization is imperative. This paper addresses this problem through the use of Dynamic Control System that Changes the number of server machines dynamically based on consumers' waiting time. A comparative analysis of this model is done with fixed severs based on cost. Two issues are addressed in this paper: the first is to determine which of the two models provide optimal solution in terms of server management and the second is the profitability concept of the experiment in the model that produces optimal solution. The results reveal a better cost effective and a competitive Cost-Benefit Ratio (CBR) of DCS over the fixed servers.