KCube: A novel architecture for interconnection networks

  • Authors:
  • Deke Guo;Hanhua Chen;Yuan He;Hai Jin;Chao Chen;Honghui Chen;Zhen Shu;Guangqi Huang

  • Affiliations:
  • Key Laboratory of Science and Technology for C4ISR Technology, School of Information System and Management, National University of Defense Technology, Changsha Hunan 410073, China;School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan Hubei 430074, China;Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Hong Kong, China;School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan Hubei 430074, China;Key Laboratory of Science and Technology for C4ISR Technology, School of Information System and Management, National University of Defense Technology, Changsha Hunan 410073, China;Key Laboratory of Science and Technology for C4ISR Technology, School of Information System and Management, National University of Defense Technology, Changsha Hunan 410073, China;Key Laboratory of Science and Technology for C4ISR Technology, School of Information System and Management, National University of Defense Technology, Changsha Hunan 410073, China;Key Laboratory of Science and Technology for C4ISR Technology, School of Information System and Management, National University of Defense Technology, Changsha Hunan 410073, China

  • Venue:
  • Information Processing Letters
  • Year:
  • 2010

Quantified Score

Hi-index 0.89

Visualization

Abstract

This paper proposes a novel architecture called KCube. KCube is a compound graph of Kautz digraph and hypercube. It employs the hypercube topology as a unit cluster and connects many such clusters by means of a Kautz digraph. It then utilizes the topological properties of hypercube to realize convenient embedding of parallel algorithms, and the short diameter of Kautz graph to support efficient inter-cluster communication. KCube possesses many attractive characteristics, such as modularity, expansibility, and regularity, while these benefits are achieved at the cost of only increasing the degree of any node by one, regardless of the network size. The methodology to construct KCube can also be applied to other compound networks after minimal modifications.