Heuristics For Virtual Machine Level Rescheduling For Improving QoS In Public Computing Utilities

This thesis deals with the issue of augmenting public resources with fixed size dedicated (private) resource pools to increase the probability of meeting the service level agreements (SLAs). Since service hosting systems based on virtual machines (VMs) are becoming increasingly practical, this thesis proposes an VM-based infrastructure for outsourcing computer services in public computing utility. Several service-oriented heuristics are proposed to perform resource management in migrating the VMs in order to obtain higher levels of quality of service. Through extensive simulations, this thesis evaluates the performance of the heuristics under different conditions.This thesis is organized into 5 chapters.Chapter 2 is concerned with the related work found in literature on resource management of computing utilities and the use of virtual machines.Chapter 3 defines the service-oriented service model, states the underlying assumptions and presents the scheduling heuristics this thesis uses.Chapter 4 demonstrates a thorough analysis of the heuristics using simulation results.Chapter 5 provides a complete summary of this thesis and presents directions for future related research.