Home // EMERGING 2011 , The Third International Conference on Emerging Network Intelligence // View article


Multi-Objective Optimization for Virtual Machine Migration on LANs for Opportunistic Grid Infrastructures

Authors:
Nathalia Garcés
Nicolás Ortiz
David Mendez
Yezid Donoso

Keywords: MOP; SPEA; resource sharing; Grid Computing

Abstract:
This paper illustrates how to apply a solution for a multiple objective problem in a simple and efficient way through the case study of an example where we must copy a single file, in this case a virtual machine, on to all computers of a LAN. Our solution is intended to be used for the creation of Virtual Clusters, which are clusters composed of virtual machines that execute on opportunistic grid infrastructures. We specify the restrictions through a mathematical model and then proceed to implement a two-part solution: First, we use the solver CONOPT to determine the Pareto frontier; then we implement an evolutionary algorithm to generate possible solutions, and match them to the Pareto frontier. Finally, we evaluate our solution as an efficient way of solving the problem through the result’s attributes and conclude which are the advantages of using evolutionary algorithms to find an answer for a multiple objective problem (time, possible solutions and variability between them).

Pages: 45 to 49

Copyright: Copyright (c) IARIA, 2011

Publication date: November 20, 2011

Published in: conference

ISSN: 2326-9383

ISBN: 978-1-61208-174-8

Location: Lisbon, Portugal

Dates: from November 20, 2011 to November 25, 2011