Cluster Computing

, Volume 18, Issue 3, pp 1041–1062

Agent-based load balancing in Cloud data centers

  • J. Octavio Gutierrez-Garcia
  • Adrian Ramirez-Nafarrate
Article

DOI: 10.1007/s10586-015-0460-x

Cite this article as:
Gutierrez-Garcia, J.O. & Ramirez-Nafarrate, A. Cluster Comput (2015) 18: 1041. doi:10.1007/s10586-015-0460-x

Abstract

Cloud data centers are generally composed of heterogeneous commodity servers hosting multiple virtual machines (VMs) with potentially different specifications and fluctuating resource usages. This may cause a resource usage imbalance within servers that may result in performance degradation and violations to service level agreements. This work proposes a collaborative agent-based problem solving technique capable of balancing workloads across commodity, heterogeneous servers by making use of VM live migration. The agents are endowed with (i) migration heuristics to determine which VMs should be migrated and their destination hosts, (ii) migration policies to decide when VMs should be migrated, (iii) VM acceptance policies to determine which VMs should be hosted, and (iv) front-end load balancing heuristics. The results show that agents, through autonomous and dynamic collaboration, can efficiently balance loads in a distributed manner outperforming centralized approaches with a performance comparable to commercial solutions, namely Red Hat, while migrating fewer VMs.

Keywords

Load balancing Virtual machine live migration Data centers Cloud computing  Intelligent agents Multiagent systems 

Copyright information

© Springer Science+Business Media New York 2015

Authors and Affiliations

  • J. Octavio Gutierrez-Garcia
    • 1
  • Adrian Ramirez-Nafarrate
    • 2
  1. 1.Department of Computer ScienceInstituto Tecnológico Autónomo de MéxicoMexico CityMexico
  2. 2.Department of Industrial & Operations EngineeringInstituto Tecnológico Autónomo de MéxicoMexico CityMexico

Personalised recommendations