TPRC 2024: The 52nd Research Conference on Communications, Information and Inter

Link to Paper

Abstract:
It is no longer clear who should pay whom and how much for interconnection between Internet Service Providers (ISPs) and content providers. Large ISPs claim that large content providers are imposing a cost on ISPs by sending large amounts of traffic to their customers. ISPs claim that it is more fair that content providers pay for this cost than consumers, because then this cost will be paid only by those consumers with high usage. In contrast, large content providers (including content delivery networks (CDNs)) claim that when they interconnect with ISPs at interconnection points (IXPs) close to consumers, they are already covering the costs of carrying traffic through the core network, and that consumers are already covering the costs of carrying traffic through the ISP's access network. These disputes between large ISPs and large content providers have recurred often during the last 10 years. When not resolved, large ISPs have often refused to increase capacity at interconnection points with large content providers and transit providers, resulting in sustained congestion which has degraded users' quality of experience.

In this paper, we examine internet interconnection policies to analyze the effects of paid peering arrangements between ISPs and content providers/transit providers using the two-sided market model. Our key findings indicate that paid peering is unlikely to result in lower consumer broadband prices. Our results also show that if a content provider or transit provider provides sufficient localization of exchanged traffic, an ISP incurs the same cost as it does when it agrees to settlement-free peering with another ISP. Our research also shows that the public interest is best served by peering prices that are lower than those likely charged by large ISPs.

Our analysis shows that settlement-free peering is warranted if a content provider or transit provider provides sufficient localization of exchanged traffic. Traffic is sufficiently localized if: (1) they interconnect at a reasonable number of interconnection points, (2) the locations of these interconnection points span the country, and (3) the proportion of traffic that is exchanged at an interconnection point that is relatively close to the end user is sufficiently high. In particular, our analysis shows that in the case of peering between an ISP and a content provider, settlement-free peering is warranted when they interconnect at a minimum of 6 interconnection points and localize at least 50% of the traffic.

Therefore, we argue that the Federal Communications Commission (FCC) should require ISPs to offer settlement-free peering arrangements to content and transit providers that agree to reasonably localize their exchanged traffic in this manner. We propose this policy recommendation as a means to serve the public interest by facilitating lower effective prices for internet data transport than market-negotiated paid peering fees. We also propose that the Commission should continue to monitor Internet traffic exchange arrangements under sections 201 and 202. In addition, however, it is now time for the Commission to determine that certain types of Internet traffic exchange arrangements are unreasonable or unreasonably discriminatory practices and would violate sections 201 or 202.

Link to Paper

Abstract:
Regulatory agencies have expressed concerns that usage-based pricing (UBP) of internet service steers consumers from streaming video to traditional TV subscriptions. We study this issue with household-level panel data from an internet service provider’s UBP experiment, capturing the pricing strategy’s effects on internet and TV subscriptions, application-specific internet usage, and payments to the firm. UBP served largely to meter internet usage by high-demand households rather than steering them toward TV. Households’ payments increased due to usage-related overage charges and internet subscription upgrades to avoid overages. Households that avoided internet-related payments reduced their internet usage rather than adding TV subscriptions.

Link to paper

Abstract:
The availability of reliable, high-speed internet throughout the United States has been a focus of policymakers for decades, with the need for an expansive broadband infrastructure listed as “the great infrastructure challenge” of the 21st century by the Federal Communication Commission. Federal policy has been implemented to confront this challenge, with investments targeting unserved and underserved regions of the country totaled $50 billion from 2009 – 2017, with $44 billion also spent from 2015 - 2020. Additionally, internet service providers (ISPs) have invested hundreds of billions more in broadband infrastructure, with estimates as high as $102 billion in 2022 alone.

Despite this investment, millions of Americans still need a broadband connection at home. A big reason for this lack of connection is the cost of services. Recent attempts to assist low-income households have included the Emergency Broadband Benefit (EBB) and the Affordable Connectivity Program (ACP). Additionally, Congress allocated nearly $65 billion for broadband programs through the Infrastructure Investment and Jobs Act (IIJA) of 2021, with the majority of funding directed to the Broadband Equity, Access, and Deployment Program (BEAD). A key component outlined in the program’s notice of funding opportunity is that state broadband officials, tasked with implementing the program and distributing billions of federal dollars in new investments, must ensure that “high-quality broadband services are available to all middle-class families … at reasonable prices” for BEAD-funded projects.

However, federal guidelines have never set a benchmark for what “reasonable prices” might mean for residents in each state, and there are no established benchmarks for determining what an affordable level of broadband service would look like throughout the country. Complicating the matter is the federal government’s goal of universal broadband access by 2030, pledging to connect “every resident and small business to reliable, affordable high-speed internet” (The White House, 2023). However, as with previous telecommunications policies, the lack of a precise definition for what “affordable” means hinders achieving the goal of universal, affordable access (Crandall and Waverman, 2000). To advance assessments about affordability, this paper opens the discussion on what publicly available data would be needed related to broadband pricing that could fit within a continuum approach to broadband.