Enhancing Network I/O Virtualization for Cloud Using Finite Multi Server Queuing Model

Abstract

Virtualization is an important technology that separates computing functions and technology implementation from physical hardware. It reduces costs by increasing energy efficiency and requiring less hardware. The major drawback of network I/O virtualization is its poor networking performance. In this paper the networking performance of virtual machines is calculated on the basis of priority. Also it should allow multiple applications to run simultaneously without affecting the networking performance. For this purpose we put forward a system where we use multi server queuing model where the applications are modeled as queues and the virtual machines are modeled as service providers.