e-OnTheInternetSearchAbout e-OTIArchivesThemesNewsHome

Article NavigationPrintable VersionEmail to a Friend

Quick Search of e-OTI

ISOC LogoISOC Home PageISOC Home Page

IllustrationThe International Internet Interconnection Issue

By Jane van Beelen <jane.vanbeelen@corpmail.telstra.com.au> and John Rolland <john.rolland@team.telstra.com>

The Issue

Transmission of Internet traffic between countries is, at the highest level, that there is neither transparent nor uniform as to how the costs of providing Internet access and Internet content services are allocated among the Internet service providers (ISPs) who incur costs in providing those services or who derive revenue from them. This issue manifests itself in a perception that the net distribution of costs and benefits arising from provision of Internet access and Internet content services does not reflect either the distribution of value or the distribution of investment.

For example, because the Internet started in the United States, Asia-Pacific ISPs-and in fact, most of the non-American ISPs-currently pay the full cost of the leased lines connecting their countries to a U.S. international gateway, together with port charges to connect to a U.S. Internet backbone. This is despite the fact that some 30 percent of the traffic on the lines is arguably attributable to U.S. Internet users accessing content located on servers in the Asia-Pacific region.

There is thus a perception that nonU.S. ISPs invest more in order to receive the same benefit as U.S. ISPs: global Internet connectivity. The result is inequitable cost allocation: Asia-Pacific ISPs pay for and receive the benefit of connectivity to the U.S., whereas U.S. ISPs receive the benefit of connectivity to the Asia-Pacific region without paying for it-and such is the case for most international destinations. Thus, Asia-Pacific ISPs perceive that U.S. ISPs leverage the investment of ISPs in other countries, so that the U.S. gets a net financial benefit that is not equivalent to the amount of service it provides in return.

The business model and technical characteristics of the Internet are such that to date, only two stable interconnection models have emerged. They are:

Sender-keep-all. This usually applies when two ISPs agree that the value1 of connecting their networks is roughly equal and they therefore interconnect on a payment-free basis. Thus, traffic may be exchanged between their networks, with each bearing its own costs.

Customer/supplier. This usually applies when the value of interconnection for each of two networks is disparate-for instance, when a small ISP connects to a larger ISP that operates a national Internet backbone. In this circumstance, the small ISP is the customer acquiring Internet connectivity from the larger ISP, and it pays for that service.

That there are only two stable interconnection models is symptomatic of the industry issue: there is no transparent and uniform model either for Internet interconnection or for Internet content provision. Interconnection on the basis of settlements to reflect disparate traffic flows is open to abuse because currently, the systems and business models necessary to implement such an arrangement do not exist. A number of organizations are trying to resolve these issues. But even if the technical issue regarding traffic flows is resolved and if some form of a carriage-cost-sharing model becomes possible, there still will be no uniform business model for Internet content.

The issue has two practical implications for Asia-Pacific ISPs: Asia-Pacific Internet users pay higher charges for international Internet connectivity than their U.S. counterparts, and Asia-Pacific Internet-based businesses have an incentive to relocate to the U.S., where global Internet connectivity is cheaper.

The Parties

Key non-U.S. ISPs concerned about the practical implications of this issue include Telstra (Australia), KDD (Japan), Dacom (South Korea), SingTel (Singapore), and Cable & Wireless HKT (Hong Kong). Industry associations representing ISPs and the information technology industry are also concerned. They include ASSOCIO, the Australian Information Industry Association, and the Internet Industry Association in Australia. The governments of Australia and of a number of Asian and European nations are also very concerned about this issue. The concern is most pronounced in the Asia-Pacific region because the costs of transpacific carriage are greater than those of transatlantic carriage. The position of these parties is that U.S. ISPs should pay at least some of the costs of carrying the international traffic generated by their customers.

Internet usage in the Asia-Pacific region is increasing in both directions due to increased e-mail and e-commerce transactions and will continue to expand. Many Asia-Pacific ISPs believe that if the Internet is to continue to grow in their region, they cannot afford to subsidize U.S. Internet users. Now, most Asia-Pacific ISPs pay the entire cost of data links to U.S. points of interconnection, in addition to port charges for interconnection to U.S. Internet backbones. They see inequity in both the lack of recognition for value brought by their extensive networks within their own economies and the lack of compensation for the costs they incur in carrying traffic generated by North Americans on their networks. As such, U.S. ISPs are perceived to be free riding on infrastructure paid for by nonU.S. ISPs. Asia-Pacific ISPs seek a more equitable arrangement in the area of cost sharing.

The attitude of U.S.-based ISPs seems to be that all ISPs-including those based outside the U.S., as well as all end users-benefit from international traffic. The U.S. ISPs argue that due to the nature of the technology-for instance, nonspecific routing of packets for delivery-it is very difficult, or even impossible, to determine the direction and the volume of the traffic. In addition to this, they also argue that due to this uncertainty and to the fact that so many parties benefit from Internet traffic, it is very difficult to ascertain the cost that traffic would incur given that the determination of price for Internet traffic is so different from that of voice telephony. Asia-Pacific ISPs essentially agree with those views. As discussed previously, the key issue is that the direction and quantum of traffic flow are not necessarily the direction and quantum of commercial benefit. The Asia-Pacific ISPs' view differs from the U.S. position because the Asia-Pacific ISPs have a pressing commercial incentive to address this issue.


The following forums have considered the issue:
  • APEC-Tel/ICAIS Task Force: In late 1998, APEC-Tel formed a special task force-the International Charging Arrangements for Internet Services (ICAIS) Task Force-to study this issue and report on it, including recommendations as to how the issue might be addressed, to the 4th Meeting of APEC Ministers for Telecommunications (TelMin-4), which will be held in Cancún, Mexico, on 24-26 May 2000.

The work of the ICAIS Task Force is culminating in the finalization of its report to TelMin-4. The report includes a set of agreed points in relation to international Internet charging arrangements that the ministers will be invited to consider endorsing at their meeting.

  • ITU Study Group 3: ITU Study Group 3 (SG3) is concerned with charging and accounting practices for all international telecommunications services. The international Internet charging issue was discussed at the meeting of the Asia and Oceania Regional SG3 Group (referred to as the TAS Group) in February 2000. The TAS Group decided to put forward to the full SG3 meeting a proposed new ITU-T Recommendation to apply to international Internet charging arrangements, discussed at the SG3 Meeting Plenary in Geneva in April 2000.

It is notable that this issue is not yet being considered by the World Trade Organization. It seems likely that ultimately, the WTO will need to address a variety of issues related to international trade in content.


Any resolution would likely be tied into resolution of international trade issues pertaining to trade in online content, which may occur if the World Trade Organization pursues that issue.

A number of industry initiatives should reduce the costs of international Internet interconnection, but no initiatives have been taken to reduce the perceived inequity in relation to who bears those costs.

Industry cost-reduction initiatives include the following:

  • Large international ISPs or aggregated groups of international ISPs could establish more points of presence (POPs) in the U.S., as already undertaken by a group of Asia-Pacific ISPs.

  • Asia-Pacific ISPs are setting up their own Internet access points in strategic areas within the region. This facilitates direct routing between sites rather than via the U.S., thereby reducing the cost of interconnection.

  • Some ISPs are investing in unidirectional international satellite capacity to supplement their fiber links.

In any event, the costs of international carriage are decreasing rapidly by orders of magnitude, which may ultimately reduce the significance of the perceived carriage cost inequity. But this does not in any way lessen the need to remove the current perceived inequity. As each day goes by, the non-U.S. Internet industry continues to believe that it is being financially disadvantaged by subsidizing the U.S. Internet industry.

Regulatory imposition of a solution is neither invited nor recommended. The lack of a uniform Internet business model means that no one-size-fits-all solution is appropriate, and there is tremendous scope for regulatory error in complex and dynamic Internet markets. There probably is no ideal solution for all parties-that's the nature of competition. But the global Internet industry will take a significant step toward resolving this issue if it can agree to adopt a fundamental, fair trading principle: that an operator using another operator's network to send traffic to and from its customers will recognize the value of being able to use that network for the conduct of its business.


1. The authors note that the assessment of value is inherent in the process of commercial negotiation.

About the Authors

The authors are Jane van Beelen and John Rolland of Telstra Corporation Limited. Telstra is one of many non-U.S. carriers that are essentially seeking wider acknowledgment of this issue. As an Australian ISP, Telstra understands the commercial drivers of ISPs and is not seeking to promote noncommercial solutions or regulatory intervention. Widespread industry acknowledgment and understanding of the issue is surely a step toward industry resolution of the issue.

Bottom Navigation BarTopSearchAbout e-OTIArchivesThemesNewsHome