1st International ICST Conference on Communications and Networking in China

Research Article

On routing controls in ISP topologies: A structural perspective

  • @INPROCEEDINGS{10.1109/CHINACOM.2006.344870,
        author={Ryota  Fukumoto and Shin’ichi Arakawa and Masayuki Murata},
        title={On routing controls in ISP topologies: A structural perspective},
        proceedings={1st International ICST Conference on Communications and Networking in China},
        publisher={IEEE},
        proceedings_a={CHINACOM},
        year={2007},
        month={4},
        keywords={},
        doi={10.1109/CHINACOM.2006.344870}
    }
    
  • Ryota Fukumoto
    Shin’ichi Arakawa
    Masayuki Murata
    Year: 2007
    On routing controls in ISP topologies: A structural perspective
    CHINACOM
    IEEE
    DOI: 10.1109/CHINACOM.2006.344870
Ryota Fukumoto1,*, Shin’ichi Arakawa1,*, Masayuki Murata1,*
  • 1: Graduate School of Information Science and Technology, Osaka University
*Contact email: r-fukumoto@ist.osaka-u.ac.jp, arakawa@ist.osaka-u.ac.jp, murata@ist.osaka-u.ac.jp

Abstract

Recent measurement studies on the Internet topology show that connectivity of nodes exhibits power-law attribute, but it is apparent that only the degree distribution does not determine the network structure, and especially true when we study the network-related control like routing control. In this paper, we evaluate the optimal routing and several routing methods on the ISP router level topologies where degree distributions follow power-law. We then examine how structural characteristics of topologies affect the network performance. The evaluation results show that the optimal routing method in a topology obtained by a modeling method considerably increases network throughput. However, the optimal routing method in ISP topologies cannot achieve high network throughput as observed in the modeling-based topology. Our results also show that our proposed routing method achieves almost the same network throughput to the optimal routing method in power-law networks, and more importantly, it exhibits the similar distribution of link utilization.