An enhanced load balancing mechanism based on deadline control on GridSim

Yongsheng Hao*, Guanfeng Liu, Na Wen

*Corresponding author for this work

Research output: Contribution to journalArticle

36 Citations (Scopus)

Abstract

A Grid is a network of computational resources that may potentially span many continents. Load balancing in a Grid is a hot research issue which affects every aspect of the Grid, including service selection and task execution. Thus, it is necessary and significant to solve the load balancing problem in a Grid. In this paper, we propose a dynamic, distributed load balancing scheme for a Grid which provides deadline control for tasks. In our scenario, first, resources check their state and make a request to the Grid Broker according to the change of load state. Then, the Grid Broker assigns Gridlets between resources and scheduling for load balancing under the deadline request. We apply our load balancing strategy into a popular Grid simulation platform GridSim. Experimental results prove that our proposed load balancing mechanism can (1) reduce the makespan, (2) improve the finished rate of the Gridlet, and (3) reduce the resubmitted time.

Original languageEnglish
Pages (from-to)657-665
Number of pages9
JournalFuture Generation Computer Systems
Volume28
Issue number4
DOIs
Publication statusPublished - Apr 2012

Keywords

  • Deadline control
  • Gridlet
  • GridSim
  • Load Balancing

Fingerprint Dive into the research topics of 'An enhanced load balancing mechanism based on deadline control on GridSim'. Together they form a unique fingerprint.

  • Cite this