The Community for Technology Leaders
Green Image
The Internet is a hierarchical architecture comprising heterogeneousentities of privately owned infrastructures, where higherlevel Internet service providers (ISPs) supply connectivity to the localISPs and charge the local ISPs for the transit services. One of thechallenging problems facing service providers today is how to increasethe profitability while maintaining good service qualities as the networkscales up. In this work, we seek to understand the fundamental issueson the "interplay" (or interaction) between ISPs at different tiers. Whilethe local ISPs (which we term peers) can communicate with each otherby purchasing the connectivity from transit ISPs, there stands an opportunityfor them to set up private peering relationships. Under this competitiveframework, we explore the issues on (a) impact of peering relationship,(b) resource distribution, (c) revenue maximization, and (d)condition for network upgrade. Firstly, a generalized model is presentedto characterize the behaviors of peers and the transit ISP, in which theireconomic interests are reflected. We study how a peer can distributivelydetermine its optimal peering strategy. Furthermore, we show how atransit ISP is able to utilize the available information to infer its optimalpricing strategy, under which a revenue maximization is achieved. Twodistributed algorithms are proposed to help ISPs to provide a fair andefficient bandwidth allocation to peers, avoiding a resource monopolizationof the market. Last but not least, we investigate the above issues ina "many-peers-region", i.e., when we scale up the network. We provideinsightful evidence to show that the ISPs can still gain profits as theyupgrade the network infrastructures. Extensive simulations are carriedout to support our theoretical claims.
ISP peering, economic pricing, distributed resource allocation, scalability.

J. W. Jiang, D. C. Chiu, J. C. Lui and S. C. Lee, "Interaction of ISPs: Distributed Resource Allocation and Revenue Maximization," in IEEE Transactions on Parallel & Distributed Systems, vol. 19, no. , pp. 204-218, 2007.
98 ms
(Ver 3.3 (11022016))