CONCLUSION
The norm of revenue-neutral peering has been breaking down for some time, and is being replaced by less constrained negotiation about paid peering, partial transit and other such variants.
Paid peering may increase the number of direct peering connections and reduce transit costs, thus reducing ISP operating cost, but it also may increase transaction costs around peering agreements. The emergence of new norms for such contracts may help ameliorate such costs.The emergence of networks that deliver high-volume commercial content (such as network CD in our discussion) raise potential concerns about the market power of both content delivery and access networks, given that there need to be payments for both content and transport flowing within the ecosystem. Interconnection between networks like our examples CD and A are a type of peering agreement insofar as routing is restricted (the traffic is destined for A’s ‘eyeball’ subscribers), but with very asymmetric (one-way) traffic flows that contribute to significant traffic/usage-related costs (Ca) for the access network. The need to recover these costs provides a reasonable economic efficiency rationale for network CD paying network A. However, if network A has sufficient market (bargaining) power, there is a risk that network A might be able to extract excessive payments from network CD, potentially even extracting content-related payments and thus leveraging its market power beyond network services.
While this remains a valid concern, the current complex mesh of ISP interconnections that include multi-connected content delivery and ISP networks bound together with a mix of peering, transit, paid-peering and partial transit interconnection agreements constrains the ability of access ISPs to over-charge content delivery networks. For example, the price of competitive transit services provides a soft upper bound for what content delivery networks might be induced to pay access networks.
Additionally, today’s content delivery networks operate thousands of servers with multiple interconnection points that provide dynamic flexibility for how these networks inject their traffic into interconnection partner networks, and thereby provide them with bargaining leverage to influence costs incurred by interconnection partner networks.To the extent access ISPs do not recover their incremental traffic costs (Ca) from upstream content delivery networks (or absorb those costs from their profits), the access ISPs may look to new usage-based pricing models to recover those costs. Anecdotal evidence on usage patterns and network costs suggests that there are significant costs that may need to be recovered, and could justifiably support the emergence of usage-based tiered pricing, but the costs are not so significant (yet) as to imply that such tiered pricing is necessary. Its emergence is likely to hinge, in part, on the implicit subsidy from light to heavy users that is implicit in flat-rate uniform pricing. However, any significant move to usage-based pricing will attract significant public attention in the USA (where consumers have become accustomed to flat-rate pricing) and will raise justifiable regulatory concern about the potential abuse of market power by access ISPs in both the retail and wholesale interconnection markets.
This poses a difficult challenge for regulatory policy-makers. Our analysis suggests that we ought to focus first on improving transparency into the workings of the Internet ecosystem in general, interconnection markets more specifically. We identified three categories of information that would contribute to improved transparency: (1) information about industry-wide cost models; (2) information about traffic trends and distributions; and (3) information about terms and conditions of interconnection agreements. Better data in these areas will contribute to the public debate, even if some categories of information are deemed too sensitive to be shared except under some restrictive confidentiality protections.
This will help provide a foundation on which a better assessments of the need for further regulations might be made; should such a need be identified, this information will also be instrumental in helping to craft suitable rules.NOTES
* The authors would like to acknowledge support from NSF Awards 1040020 and 1040023, and the MIT Communications Futures Program (http://cfp.mit.edu). All opinions expressed herein are those of the authors alone.
1. Because not everyone needs to make telephone calls at the same time, source and destination nodes can share peak capacity that only needs to be large enough to meet the demand at the traffic peak. This resource sharing, which relies on the statistical multiplexing of supply and demand, is a key feature of the network planning.
2. Many transmission (link) technologies have a significant fixed cost component that does not vary significantly with link capacity or traffic. Thus transmission costs typically scale sublinearly with capacity (scale economies).
3. A positive network externality arises when the value of adding an additional subscriber to a network is greater than just the value realized by that subscriber from joining the network. A telephone network that allows a subscriber to call anyone anywhere is more valuable to each subscriber than would be a telephone network that only allowed each subscriber to call people in a single town.
4. For general discussions of network interconnection economics, see Armstrong (2002).
5. For a discussion of the need to rationalize intercarrier compensation by adopting bill-and-keep, see DeGraba (2000).
6. An AS is the unit of administrative policy in the Internet, comprised of one or more networks that are controlled by a common network administrator. An AS might be a university, a business enterprise, or a part of a division.
7. Internet exchanges where multiple providers interconnect to exchange traffic can be used for both peering and transit agreements.
In the early days of the Internet, much of the traffic was exchanged at public interexchange points such as MAE-East and MAE-West where a large number of networks physically interconnected to exchange traffic without compensation. These early exchanges had relatively open interconnection policies and provided an inexpensive way for smaller ISPs to expand connectivity. However, they lacked mechanisms to provide incentives to provide capacity to deliver traffic from the public interexchange points. As a consequence, these public exchanges suffered from significant congestion. The bulk of Internet traffic subsequently shifted to bilateral interconnection agreements.8. See, for example, Laffont et al. (2001, 2003), Greenstein (2004) and Besen et al. (2001).
9. For further discussion of how interconnection markets have been changing, see Farratin et al. (2007).
10. For example, the typical residential broadband service offers a peak data rate in the tens of Mbps, whereas many enterprise customers have access connections offering thousands of Mbps or higher data rates.
11. For example, the current version of the Border Gateway Protocol (BGP) used to route traffic between AS is Version 4, adopted in 2006, which updated and added functionality to earlier versions of BGP.
12. This estimate is for residential traffic on fixed access networks in North America during the first half of 2013 (see Sandvine, 2013).
13. Since the time of writing, Netflix appears to have changed its content delivery approach, but it is not clear precisely what it is doing. It is to be expected that such changes may occur.
14. Basic telecommunication services are regulated as common carrier services under Title II of the United States Communications Act of 1934 (as amended). Over time, the Federal Communications Commission (FCC) has grappled with defining the boundary between what services should or should not be regulated as basic telecommunication services (see, e.g., Brock, 1994).
In a succession of decisions since 2002, the FCC sought to reclassify broadband services as ‘information services’ rather than basic telecommunication services. By so doing, they opened the possibility of shifting to a more light-handed form of regulation that could be more technology neutral, and evolve beyond the silo-based legacy regulatory models that characterized traditional telecommunications and cable television regulation. For a discussion of evolving broadband policy, see ‘Notice of Proposed Rulemaking, In the Matter of Preserving the Open Internet and Broadband Industry Practices’, Federal Communications Commission, GN Docket No. 09-191 and WC Docket No. 07-52, released 22 October 2009 (hereafter, NPRM, 2009).15. First Amendment ‘free speech’ concerns induce a special level of concern for content-based regulation in the USA, providing additional motivation for separating content and conduit regulation as much as possible. Traditional content regulation includes such things as program access rules (e.g., to ensure that incumbents make programming available to other distribution channels), public programming obligations (e.g., children and news programming), and censorship rules (e.g., pornography restrictions).
16. For example, a significant goal of the Telecommunications Act of 1996, which embodied significant reforms to the Communications Act of 1934, was to significantly expand the scope of competition in all telecommunication services and provide a roadmap for further deregulation. In the years that followed, talk turned to notions of layered regulation that would be more appropriate for the emerging world of facilities-based competition between multi-service platform networks offered by telephone, cable television, and potential new entrants (see, e.g., Werbach, 2002; Friedan, 2004; Sicker et al., 2007).
17. See, for example, Jordan (2007) who relates the layered regulation and network neutrality debates; or Lehr et al. (2007).
18. For example, some have argued that access ISPs have a terminating monopoly because subscribers face switching costs.
This source of market power might exist even if the market for selecting a broadband service provider is competitive (see NPRM, 2009, note 15 supra at paragraph 73). For a critique of this view, see ‘Testimony of Jeffrey Eisenach, PhD, Before the Subcommittee on Communications, Technology and the Internet, Committee on Energy and Commerce, United States House of Representatives, 21 April 2010’.19. Content-specific costs include the costs of producing and promoting the content, and are logically separable from the costs of distributing the content via the Internet. One might also imagine there is a value to consumers in viewing the content, irrespective of the medium via which the content is viewed (in a movie theater vs at home).
20. The separation is not precise because the way in which content is delivered may impact its quality, and hence its value to the end user. For example, high-definition content may be down-coded for delivery to the small screen without deterioration in the user experience; alternatively, increased latency or congestion may severely adversely impact the user experience.
21. See ‘ESPN charges broadband firms for access to ESPN3.com content’, Investors.com, 24 August 2010, accessed 15 April 2015 at http://news.investors.com/technology/082410-544885-espn-charges-broadband- firms-for-access-to-espn3com-site.htm.
22. We see no evidence from their published annual reports that content delivery networks are earning significant supernormal profits. Moreover, anecdotal evidence suggests that the prices for content delivery networks have followed the downward trend in transit pricing, and are quite modest. See, for example, ‘Internet transit pricing-historical and projected’, accessed 15 April 2015 at http://drpeering. net/white-papers/Internet-Transit-Pricing-Historical-And-Projected.php, DrPeering International, 2010, and ‘Data from Q1 shows video CDN pricing stabilizing, down 25% in 2010’, Streamingmedia.com, 2010, accessed 15 April 2015 at http://blog.streamingmedia.com/the_business_of_online_vi/2010/06/data-from- q1-shows-video-cdn-pricing-stabilizing-should-be-down-25-for-the-year.html.
23. To the extent either may occur, we assume that these flow either via the content-producer/owner relationship to the content delivery networks or via the subscriber relationship to the access ISP.
24. This measurement in terms of cost/GB is perhaps confusing. As written, it is not a rate but a volume. That is, the implication is that it costs the ISP (say) $0.10 to deliver a GB of data, independent of rate. This characterization is obviously a simplification, but it implies that the cost to deliver a GB in one unit of time at one rate, or the cost to deliver that same GB at half the rate over twice the time is more or less the same. Another way of saying this is that $0.10/GB is a contraction of ‘$0.10/month for each GB/month’.
25. An estimate of CAD0.08/GB is given in ‘What does a gig cost?’, Michael Geist, 6 April, accessed 15 April 2015 at http://www.michaelgeist.ca/content/view/5727/125/ and ‘What does a gigabyte cost, revisited’, Michael Geist, 29 July, accessed 15 April 2015 at http://www.michaelgeist.ca/content/view/5952/125/.
26. In contrast, these numbers will not apply to rural ISPs that may be far from peering and transit interconnection points, or smaller networks, wherever they are located.
27. See ‘Re: Ex parte, GN Docket No. 09-191, WC Docket No. 07-52’, accessed 15 April 2015 at https:// prodnet.www.neca.org/publicationsdocs/wwpdf/051211netflix.pdf.
28. Usage costs are expected to grow sublinearly with traffic. Also, median per customer usage has grown even faster, from 6.0 GB (2011) to 16.0 GB (2013). The high ratio of mean-to-median indicates a heavy-tailed distribution, with some users being very heavy indeed (see Sandvine, 2011, 2013).
29. See Sirbu and Agyapong (2011) for an alternate model of the relationship between content delivery networks and access ISPs that focuses on how routing/interconnection choices might influence total costs and potential payment flows.
30. See Dhamdhere et al. (2010) for an interesting alternative model for interconnection.
31. We noted the case above of ESPN3, which had content of sufficient popularity that it bargained with access networks such as Comcast to pay ESPN3 a per customer fee, which presumably is then passed on to the customers.
32. See ‘Internet peering, paid peering and Internet transit’, DrPeering International, 14 January, accessed 15 April2015athttp://drpeering.net/AskDrPeering/blog/articles/Ask_DrPeering/Entries/2011/1/14_Internet_ Peering,_Paid_Peering_and_Internet_Transit.html.
33. The internal cost C0 of this service is low: only the load on the router backplane passing the traffic from one port to another. Hence the name ‘single-hop’; this configuration has also been called ‘backplane access’.
34. In the past, such control might have been limited to choosing between hot or cold potato routing. In hot (cold) potato routing the source ISP hands over the traffic sooner (later), thereby maximizing (minimizing) use of the destination ISP’s resources. An Alcatel-Lucent white paper argues that the rise of large and increasingly complex content delivery networks may allow them to suck transit revenues out of the Internet ecosystem (see Alcatel-Lucent, 2011).
35. See ‘A study of 28 peering policies’, DrPeering International, accessed 15 April 2015 at http://drpeering. net/white-papers/Peering-Policies/A-Study-of-28-Peering-Policies.html.
36. Nine of the 28 peering agreements included a requirement for traffic ratios. ISPs with traffic ratio requirements include AboveNet, Comcast, Verizon, ATT, CableVision and Quest. Several of these are what we classify as access networks, which supports the hypothesis that these networks are especially concerned with how peering agreements with content delivery networks are negotiated.
37. Traffic that exceeds a tier’s allotted volume may be priced at a pre-specified coverage rate, or the customer may be temporarily boosted into a higher-volume tier, or the traffic might be shaped (subject to reduced data rates) or even dropped.
38. Zero rating of select services is deemed contentious. Some promote it as a way to lower the price for broadband and expanding access, while others view it as a mechanism for access providers to extract revenues from content providers. For a discussion of the debate, see, for example, ‘Zero rating: the FCC’s war on affordable broadband’, TechPolicyDaily, 26 May 2010, accessed 31 May 2015 at http://www.techpolicydaily.com/internet/fccs-war-on-zero-rating/ or ‘Why “zero rating” is the new battleground in net neutrality debate’, CBC News, 7 April 2015, accessed 31 May 2015 at http://www.cbc.ca/news/business/ why-zero-rating-is-the-new-battleground-in-net-neutrality-debate-1.3015070.
REFERENCES
Alcatel-Lucent (2011), ‘Analysis: Content peering and the Internet economy’, Alcatel-Lucent white paper, 19 April 2011, accessed 15 April 2015 at http://www2.alcatel-lucent.com/blogs/techzine/2011/ analysis-content-peering-and-the-internet-economy/.
Armstrong, M. (2002), ‘The theory of access pricing and interconnection’, in M. Cave, S.K. Majumdar and I. Vogelsang (eds), Handbook of Telecommunications Economics, Volume I, London and New York: Elsevier.
Besen, S., P. Milgrom, B. Mitchell and P. Srinagesh (2001), ‘Advances in routing technologies and Internet peering agreements’, American Economic Review, 91 (2), 292-6.
Brock, G. (1994), Telecommunications Policy for the Information Age: From Monopoly to Competition, Cambridge, MA: Harvard University Press.
Clark, D. (2008), ‘A simple cost model for broadband access: What will video cost?’, paper presented at the 36th Research Conference on Communication, Information, and Internet Policy, September, Arlington, VA: George Mason University.
DeGraba, P. (2000), ‘Bill and keep at the central office as the efficient interconnection regime’, OPP Working Paper Series, No. 33, Washington, DC: Federal Communications Commission, accessed 15 April 2015 at http://www.fcc.gov/working-papers/bill-and-keep-central-office-efficient-interconnection-regime.
Dhamdhere, A., C. Dovrolis and P. Francois (2010), ‘A value-based framework for Internet peering agreements’, in Proceedings of 22ndInternational Teletraffic Congress (ITC22), 7-9 September 2010, Amsterdam: IEEE, 1-8.
Faratin, P., D. Clark and P. Gilmore et al. (2007), ‘Complexity of internet interconnections: Technology, incentives and implications for policy’, paper presented at the 35th Research Conference on Communication, Information, and Internet Policy, September, Arlington, VA: George Mason University, accessed 15 April 2015 at http://ssrn.com/abstract=2115242.
Frieden, R. (2004), ‘Adjusting the horizontal and vertical in telecommunications regulation: A comparison of the traditional and a new layered approach’, Federal Communication Law Journal, 55 (207), 207-50.
Greenstein, S. (2004), ‘The economic geography of Internet infrastructure in the United States’, CSIO Working Paper No. 0046, Center for the Study of Industrial Organization, Northwestern University.
Jordan, S. (2007), ‘A layered network approach to net neutrality’, International Journal of Communication, 1, 427-60.
Laffont, J.-J., S. Marcus, P. Rey and J. Tirole (2001), ‘Internet peering’, The American Economic Review, 91 (2), 287-91.
Laffont, J.-J., S. Marcus, P. Rey and J. Tirole (2003), ‘Internet interconnection and the off-net pricing principle’, RAND Journal of Economics, 34 (2), 370-90.
Lehr, W., J. Peha and S. Wilkie (eds) (2007), ‘Special issue on network neutrality’, International Journal of Communication, 1, accessed 27 December 2015 at http://ijoc.org/ojs/index.php/ijoc/issue/view/1.
Sandvine (2011), Global Internet Phenomena Report, Spring 2011, accessed 15 April 2015 at: http://www.sand vine.com/news/global_broadband_trends.asp.
Sandvine (2013), Global Internet Phenomena Report, accessed 15 April 2015 at http://www.sandvine.com/news/ global_broadband_trends.asp.
Sicker, D., J. Mindel and C. Cooper (2007), The Internet Interconnection Conundrum: A Technology Policy Analysis of Service Provider Interconnection Issues, Bloomington, IN: iUniverse.
Sirbu, M. and P. Agyapong (2011), ‘Economic incentives in content-centric networking: A network operator’s perspective’, in Proceedings of the 39th Research Conference on Communications, Information, and Internet Policy (TPRC), George Mason University, Arlington, VA, September 2011, accessed 15 April 2015 at http:// ssrn.com/abstract=1979922.
Werbach, K. (2002), ‘A layered model for Internet policy’, Journal of Telecommunications and High Technology Law, 1, 58-64.