Affiliations: TEMAP, Department of Industrial Engineering, School of
Engineering, Seoul National University, South-Korea School of Information
Technology, International University, Bruchsal, Germany | School of Information Technology, International
University, Bruchsal, Germany
Abstract: The Internet service provider market is very competitive. Small and
medium-size Internet service providers (ISPs) are competing for customers,
while, at the same time, they are under price pressure from upstream providers.
Therefore, these ISPs have to reduce their overall cost of interconnection. In
order to address this issue, Internet exchange points (IXPs) have been built up
in recent years, which allow small and medium-size ISPs to go into public or
private peering with other ISPs. However, those ISPs do not have sufficient
information to select the optimal set of ISPs, with which they should go into
private peering agreements. In this paper, we describe an approach, which
provides ISPs with the information about the most economical interconnections
to other ISPs. This approach helps small and medium-size Internet service
providers to reduce their interconnection costs for upstream connectivity and
to improve network performance for their customers. To achieve that, our
approach uses Internet topology information in close neighborhood of the ISP
(which is determined by the set of ISPs connected to the IXP), measurement
information about the number of bytes transmitted, and traffic pricing schemes.
Based on real data, our analysis results demonstrate that our approach provides
the necessary information to ISPs for locally optimizing their interconnection
agreements (e.g. peering, sibling, transit agreements).
Keywords: ISP interconnection, peering agreements, transit agreements, traffic measurements, business intelligence, traffic analysis, routing inefficiencies, pricing, autonomous systems
