TY - GEN
T1 - Performance evaluation of a method to reduce inter-ISP transit cost caused by overlay routing
AU - Matsuda, Kazuhito
AU - Hasegawa, Go
AU - Kamei, Satoshi
AU - Murata, Masayuki
PY - 2010
Y1 - 2010
N2 - Overlay routing is an application-level routing mechanism and existing research has revealed that overlay routing can improve user-perceived performance. On the other hand, overlay routing may harm the ISPs' cost structure because of the policy mismatch between IP routing and overlay routing. In a previous study, we proposed a method to reduce inter-ISP transit cost caused by overlay routing while maintaining the effectiveness of the overlay routing. However, we evaluated the proposed method only in the PlanetLab environment, the node location of which is biased to North America. In addition, the previous evaluation focused only on limiting the degree of increase in the inter-ISP transit cost and did not explicitly consider the performance of the overlay routing. In the present paper, we evaluate the proposed method in more general network environments, which are the generalized PlanetLab environment and the Japanese commercial ISPs network environment. We also evaluate the performance of the proposed method in terms of the trade-off relationships between the overlay routing performance and the inter-ISP transit cost over the entire network. In addition, we discuss the differences in the network properties in both environments, which affect the performance of the proposed method.
AB - Overlay routing is an application-level routing mechanism and existing research has revealed that overlay routing can improve user-perceived performance. On the other hand, overlay routing may harm the ISPs' cost structure because of the policy mismatch between IP routing and overlay routing. In a previous study, we proposed a method to reduce inter-ISP transit cost caused by overlay routing while maintaining the effectiveness of the overlay routing. However, we evaluated the proposed method only in the PlanetLab environment, the node location of which is biased to North America. In addition, the previous evaluation focused only on limiting the degree of increase in the inter-ISP transit cost and did not explicitly consider the performance of the overlay routing. In the present paper, we evaluate the proposed method in more general network environments, which are the generalized PlanetLab environment and the Japanese commercial ISPs network environment. We also evaluate the performance of the proposed method in terms of the trade-off relationships between the overlay routing performance and the inter-ISP transit cost over the entire network. In addition, we discuss the differences in the network properties in both environments, which affect the performance of the proposed method.
UR - http://www.scopus.com/inward/record.url?scp=78650246841&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78650246841&partnerID=8YFLogxK
U2 - 10.1109/NETWKS.2010.5624928
DO - 10.1109/NETWKS.2010.5624928
M3 - Conference contribution
AN - SCOPUS:78650246841
SN - 9781424467051
T3 - Proceedings of 2010 14th International Telecommunications Network Strategy and Planning Symposium, Networks 2010
BT - Proceedings of 2010 14th International Telecommunications Network Strategy and Planning Symposium, Networks 2010
T2 - 2010 14th International Telecommunications Network Strategy and Planning Symposium, Networks 2010
Y2 - 27 September 2010 through 30 September 2010
ER -