rHALB: a new load-balanced routing algorithm for k-ary n-cube networks

  • Authors:
  • Huaxi Gu;Jie Zhang;Kun Wang;Changshan Wang

  • Affiliations:
  • State key lab of ISN, Xidian University, Xi'an, China;Dept. of Computing and Information Systems, University of Bedfordshire, UK;School of Computer Science, Xidian University, Xi'an, China;School of Computer Science, Xidian University, Xi'an, China

  • Venue:
  • APPT'07 Proceedings of the 7th international conference on Advanced parallel processing technologies
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

A new load-balanced routing algorithm for k-ary n-cube networks is proposed in this paper. The new algorithm, called rHALB (r Hop Aware Load-Balanced) routing algorithm, can efficiently balance traffic loads in k-ary ncube networks. It is based on the limited global information, which is a tradeoff between local and global information. rHALB makes use of the number of potential deadlocked packets to detect the congestion, and can direct the traffic to bypass any detected hotspots, thus balancing the traffic. Simulations are carried out on 2D torus networks by using OPNET. The results show that rHALB achieves a better performance (latency and throughput) under various traffic patterns than dimension order routing and Duato's algorithm, which are popularly used in commercial products, and ROMM, ACQ and GAL, which were proposed in literature recently.