On the Correlation of Geographic and Network Proximity at Internet Edges and its Implications for Mobile Unicast and Multicast Routing

Abstract

Significant effort has been invested recently to accelerate handover operations in a next generation mobile Internet. Corresponding works for developing efficient mobile multicast management are emergent. Both problems simultaneously expose routing complexity between subsequent points of attachment as a characteristic parameter for handover performance in access networks. As continuous mobility handovers necessarily occur between access routers located in geographic vicinity, this paper investigates on the hypothesis that geographically adjacent edge networks attain a reduced network distances as compared to arbitrary Internet nodes. We therefore evaluate and analyze edge distance distributions in various regions for clustered IP ranges on their geographic location such as a city. We use traceroute to collect packet forwarding path and round-trip-time of each intermediate node to scan-wise derive an upper bound of the node distances. Results of different scanning origins are compared to obtain the best estimation of network distance of each pair. Our results are compared with corresponding analysis of CAIDA Skitter data, overall leading to fairly stable, reproducible edge distance distributions. As a first conclusion on expected impact on handover performance measures, our results indicate a general optimum for handover anticipation time in 802.11 networks of 25 ms.