Margit Vanberg, Terms and conditions of Internet interconnection today in:

Margit Vanberg

Competition and Cooperation Among Internet Service Providers, page 56 - 61

A Network Economic Analysis

1. Edition 2009, ISBN print: 978-3-8329-4163-5, ISBN online: 978-3-8452-1290-6

Series: Freiburger Studien zur Netzökonomie, vol. 14

Bibliographic information
56 professional and individual membership. These organizations develop standards in open working groups and rely on consensus building for the final specifications.51 3.5 History and development of interconnection agreements The earliest computer networks, such as the ARPANET, were government-funded. Users at ARPANET sites covered only the telephone charges for dial-up connections to the network, but no fees towards the network costs. University computer networks received seed funding from the NSF, but had the obligation to become self-supporting within a reasonable time frame. Organizations participating in university networks therefore paid usage charges and membership fees, with industrial users paying a multiple of what academics paid (Jennings et al., 1986: 946). However, interconnection (the reciprocal carrying of each others traffic) between early computer networks, all operating on a non-for-profit basis, was generally on a freeof-charge basis. With the advent of commercial Internet access services, the general tradition of free interconnection between independent networks began to change. ISPs started making free traffic exchange contingent on specific requirements that needed to be fulfilled by the interconnection partner. When these requirements were not fulfilled, one interconnection party paid the other party for transit services. The following section illustrates the type of commercial interconnection agreements typical today. 3.5.1 Terms and conditions of Internet interconnection today As mentioned above, interconnection often takes place at multilateral interconnection points where several ISPs meet and can realize physical network interconnections with a number of partners. In the early commercial era of the Internet most network interconnections were realized at the NSF-designed Network Access Points (NAPs). Relatively quickly, however, the quality of interconnection at these public NAPs suffered from the exponential growth of Internet traffic (European Commission, 1998: 7). Private multilateral interconnection points, so-called Commercial Internet Exchanges (CIX) were therefore added to the Internet infrastructure. An important benefit of Internet Exchanges is the fact that many interconnections can be entered into at one geographic location, such that the costs of realizing physical interconnections with several networks are significantly reduced. Interconnection at these exchanges is still subject to bilateral contracts between the connecting parties. Operators of Internet exchanges can, however, assist in the interconnection by offering a standard contract that can be used by members that enter an interconnec- 51 See for details on the organization of the Internet Society and on the standardization process; site last visited on Feb. 15, 2008. 57 tion agreement. The operators of the exchange further assist in establishing contacts between its members, for instance, by maintaining a mailing list and distributing information of common interest. Peering Private interconnections either take the form of “peering” or “transit.” Peering contracts are characterized by the following features (see also Laffont et al., 2001: 287): (1) Peering partners agree to reciprocally carry data packages between their direct customers as well as their customers’ customers. Customers in this second sense are the end-users of ISPs that have purchased IP transit (see below) from one of the peering partners (European Commission, 1998: 7f.). Peering agreements do not cover traffic to or from third networks with which the peering parties also have peering agreements. (2) Peering is generally free-of-charge to the peering partners. Interconnection by peering most closely corresponds to the interconnection agreements of the early Internet, which was dominated by the scientific community. For these networks the assumption that both partners had roughly the same costs for carrying each others traffic was perhaps generally accurate. But more importantly, network operators in these institutional settings (universities, ministries) were not pursuing commercial interests, i.e. interconnection charges played no important role in this environment. With the growing commercialization of the Internet, it is now customary that peering is agreed to only between parties with approximately the same benefit from the traffic exchange. The peering parties must generally have comparable network coverage, comparable traffic flows, and guarantee a certain quality of service. Otherwise the parties may negotiate paid peering that includes a compensation for traffic imbalances. A further possibility of adapting the peering contract is to limit the address space covered by the peering agreement so that traffic is roughly balanced and network coverage similar.52 Transit When the discrepancy between the utility the interconnecting parties receive from entering into a peering agreement is too large, networks can instead enter into a transit relationship. The main features of a transit agreement are: (1) the transit taker pays the transit giver for delivering its traffic to and from the entire Internet. The transit giver handles the traffic of the transit taker just like traffic originating from or terminating with one of its direct customers; (2) the price of a transit agreement 52 An example for this is the adjustment Cogent and Level3, both viewed as Tier-1 carriers, made to their peering contract. Level3 had initially threatened to de-peer, arguing that Cogent profited more from the arrangement. The contract was finally renewed, however, with stricter obligations concerning the volume and type of traffic accepted under the agreement (see Networld article from October 28th, 2005). 58 generally depends on the bandwidth of the connection, sometimes also on the amount of traffic transmitted over the interconnection interface.53 Both peering and transit agreements cover parameters such as the number and geographic location of the interconnection points as well as the minimal bandwidth with which the parties are connected to these exchange points. The costs of realizing the physical network interconnection between the networks are shared depending on whether interconnection partners exchange their traffic via peering or transit agreements. Generally, ISPs will bear the cost of bringing their network to the Internet exchange point irrespective of whether they are transit customers or peering partners. In the case of peering, the costs of the link connecting the networks at the interconnection site are shared according to commercial negotiations. In the case of transit, the costs of the interconnection link are often covered by the transit-customer (European Commission, 2000: 9). The characteristics of peering and transit are summarized in Table 3.1. Table 3.1: Characteristics of peering and transit agreements 53 Interview with Scott Marcus (senior advisor for Internet technology at the Federal Communications Commission from July 2001 to July 2005) on September 25, 2003 at the FCC, Washington D.C.. Peering Transit Type of traffic accepted Traffic originating with customers of one peering party and terminating with customers of the other peering party Agreement extends also to traffic originating in third party networks and terminating in third party networks Charges for traffic exchanged Most often settlement free traffic exchange Transit customer pays for interconnection Cost of the interconnection link Shared according to commercial negotiations Generally financed by transit customer 59 Examples of peering and transit The following stylized examples will help to illustrate the significant difference in the reachability that an interconnection agreement offers depending on whether it is a transit or a peering agreement. In the case illustrated in Figure 3.5, both ISP A and ISP C have a peering agreement with ISP B. Since neither carrier has a transit agreement with ISP B, traffic not destined for a customer of ISP B’s network cannot be sent via the interconnection link by either ISP A or ISP C. As peering partners of ISP B, ISP A and ISP C are not considered customers of ISP B. In order for ISP A to be able to offer its users connectivity with network users of ISP C, ISP A therefore needs to either enter into a peering or transit agreement with ISP C directly or purchase transit from ISP B. Figure 3.5: Traffic exchange among peering partners A transit agreement with ISP B would therefore offer wider reachability to ISP A than the peering contract, because ISP A is then considered a customer of ISP B. As the European Commission has argued, “the purchase of a transit service could be more accurately described as a right on the part of an ISP to have his traffic treated as the traffic of the transit provider’s network for the purpose of exchange across a peering interface” (European Commission, 1998: 8). This is illustrated in Figure 3.6. ISP A ISP B ISP C Peering IXP IXP IXP Internet eXchangePoint 60 In Figure 3.6 ISP A has entered a transit agreement with ISP B and is therefore considered a customer of ISP B with respect to the peering agreement between ISP B and ISP C. ISP A can send traffic destined for end-users of ISP C over the exchange with ISP B. As a transit taker, ISP D is considered a customer of ISP C. The peering agreement between ISP C and ISP B will therefore also include traffic originating or terminating in ISP D’s network. In consequence, ISP A can reach the networks of ISP B, ISP C, and ISP D via its transit agreement with ISP B. In Figure 3.6 both ISP B and ISP C need only their peering agreement to have full reachability of all customers connected to the ISPs A, B, C, and D. Figure 3.6: Traffic exchange among transit and peering partners The following conclusions can be drawn from the reachability offered by peering and transit interconnection agreements: (1) ISPs in the lower and middle network levels of the Internet hierarchy may have several interconnection agreements, but they must always have at least one transit agreement, such that a default route can be pointed towards a transit provider that guarantees universal connectivity. (2) Tier-1 ISPs must have peering agreements with all other top-level ISPs in order to be able to guarantee universal connectivity. This is so because peering partners accept traffic only when it is directed to one of their direct subscribers or wholesale customers ISP A ISP B ISP C Peering IXP IXP Internet eXchange Point Transit IXP ISP D IXP 61 (lower-level ISPs that purchased transit). Traffic destined to a Tier-1 network with which no peering contract exists is not covered by the peering contracts with the other Tier-1 ISPs. The ISP would only be allowed to send this traffic via a third Tier-1 ISP if it had a transit agreement with this ISP. Having a transit agreement with a Tier-1 ISP, however, means that, by definition, the ISP is not a Tier-1 ISP. 3.6 Conclusions This chapter reviews the history of the Internet with a special focus on the development of network interconnection between ISPs. The three-tier hierarchy of the Internet is explained. Furthermore, both the technical and the commercial terms for Internet interconnection are reviewed. In chapter 2 it was stated that all ISPs can be considered essentially equal in the sense that all are active on the logical layer of Internet services provision, and all require either cooperations with firms active on the other layers of Internet services provision or need to vertically integrate into these layers themselves in order to be able to offer services to end-users. This chapter introduced the differences between ISPs active on the Tier-1 level of the Internet hierarchy to ISPs active on lower levels. Upon closer examination it was shown that the services offered on the logical layer of Internet services provision are different depending on whether one is considering a Tier-1 ISP or a lower-level ISP. Only Tier-1 ISPs have the ability to offer universal connectivity on their own account, without resorting to a transit provider. All other ISPs need to purchase transit services from at least one Tier-1 ISP, either directly or indirectly through intermediate ISPs so as to be able to guarantee universal connectivity in their network. Universal connectivity is a quality characteristic which has become essential to Internet users. One relevant question for the competition analysis to follow is therefore whether the fact that only Tier-1 ISPs can offer universal connectivity on their own account lends market power to these ISPs? This question will be at the center of the analysis of the effects of network externalities on competition, which follows in chapters 6 and 7.

Chapter Preview



Die Konvergenz der Netztechnologien, die dem Internet, der Telekommunikation und dem Kabelfernsehen zu Grunde liegen, wird die Regulierung dieser Märkte grundlegend verändern. In den sogenannten Next Generation Networks werden auch Sprache und Fernsehinhalte über die IP-Technologie des Internets transportiert. Mit den Methoden der angewandten Mikroökonomie untersucht die vorliegende Arbeit, ob eine ex-ante sektorspezifische Regulierung auf den Märkten für Internetdienste wettbewerbsökonomisch begründet ist. Im Mittelpunkt der Analyse stehen die Größen- und Verbundvorteile, die beim Aufbau von Netzinfrastrukturen entstehen, sowie die Netzexternalitäten, die im Internet eine bedeutende Rolle spielen. Die Autorin kommt zu dem Ergebnis, dass in den Kernmärkten der Internet Service Provider keine monopolistischen Engpassbereiche vorliegen, welche eine sektor-spezifische Regulierung notwendig machen würden. Der funktionsfähige Wettbewerb zwischen den ISP setzt jedoch regulierten, diskriminierungsfreien Zugang zu den verbleibenden monopolistischen Engpassbereichen im vorgelagerten Markt für lokale Netzinfrastruktur voraus. Die Untersuchung zeigt den notwendigen Regulierungsumfang in der Internet-Peripherie auf und vergleicht diesen mit der aktuellen Regulierungspraxis auf den Telekommunikationsmärkten in den Vereinigten Staaten und in Europa. Sie richtet sich sowohl an die Praxis (Netzbetreiber, Regulierer und Kartellämter) als auch an die Wissenschaft.