Volume 94, Number 2, April 2011
|Number of page(s)
|Interdisciplinary Physics and Related Areas of Science and Technology
|18 April 2011
Self-organizing weights for Internet AS-graphs and surprisingly simple routing metrics
Aarhus School of Engineering and Department of Mathematical Sciences, Aarhus University - Ny Munkegade 118, 8000 Aarhus C, Denmark, EU
2 Frankfurt Institute for Advanced Studies and Frankfurt International Graduate School for Science, Johann Wolfgang Goethe Universität - Ruth-Moufang-Straße 1, 60438 Frankfurt am Main, Germany, EU
Accepted: 17 March 2011
The transport capacity of Internet-like communication networks and hence their efficiency may be improved by a factor of 5–10 through the use of highly optimized routing metrics, as demonstrated previously. The numerical determination of such routing metrics can be computationally demanding to an extent that prohibits both investigation of and application to very large networks. In an attempt to find a numerically less expensive way of constructing a metric with a comparable performance increase, we propose a local, self-organizing iteration scheme and find two surprisingly simple and efficient metrics. The new metrics have negligible computational cost and result in an approximately 5-fold performance increase, providing distinguished competitiveness with the computationally costly counterparts. They are applicable to very large networks and easy to implement in today's Internet routing protocol on the AS-level.
PACS: 89.20.Hh – World Wide Web, Internet / 89.75.Hc – Networks and genealogical trees / 89.20.-a – Interdisciplinary applications of physics
© EPLA, 2011
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.