Interconnection and Capacity Allocation for All-IP Networks: Walled Gardens or Full Integration?

Abstract

The migration towards all-IP networks fundamentally changes the equation for IP interconnection. A wide spectrum of different application services requires heterogeneous levels of quality of service (QoS). Although the relevance of end-to-end inter-provider QoS has been recognized for a long time and technical means for implementing such strategies were developed more than a decade ago, there is no widespread implementation of differentiated IP interconnection. Taking into account the evolution in the Internet ecosystem and the transition towards all-IP, the aim of this paper is to describe the evolution and future challenges in the markets for “all-IP interconnection”. The role of content delivery networks (CDNs) for the provision of end-to-end QoS is discussed and implications for QoS-based interconnection agreements are derived. We compare a scenario based on fully integrated all-IP service provision to a scenario based on a regulatory market split between two types of traffic services as proposed by the recent FCC regulation in the U.S. and what is similarly envisaged in Europe. Based on a network economic analysis, we derive implications for an economically desirable all-IP interconnection regime. Economically optimal capacity allocation requires market driven price and QoS differentiations based on the opportunity costs of network usage. Moreover, the integrated optimization of capacity allocation should be based on an unrestricted evolutionary search for bilateral and multilateral interconnection agreements. A fully integrated solution results in an all-IP ecosystem relying on a continuum of interconnection agreements capable of reflecting heterogeneity in demand for QoS.

Keywords: all-IP, interconnection, QoS, content distributions networks, peering, transit

JEL Classification: L 51, L 86, L 96