A European proposal that threatens the Internet’s architecture
The European Commission is weighing new rules on how networks interconnect, and the implications go well beyond a commercial spat between US tech giants and Europe’s incumbent telecom operators. The stated goal—universal Gigabit connectivity across the EU by 2030—is laudable. But the accompanying exploratory consultation points toward something more specific: regulatory intervention in the Internet interconnection market designed to extract payments from large content and application providers (CAPs) and direct them to incumbent telcos.
That framing misses what is at stake. If adopted, these “network usage fees” would not just reprice Big Tech’s access to European networks. They would create fast lanes for the largest content providers and push everything else—including content from small businesses, startups, and European consumers—into congested slow lanes. The wider risk is that a regulatory fix for a problem that doesn’t exist ends up degrading the Internet for everyone.
How traffic actually reaches users
Understanding the interconnection market is essential to assessing the proposal. When a user requests content, the data travels from a content or service provider to the user’s ISP through one of three paths: a direct settlement-free peering connection, a connection at an Internet exchange point (IXP), or via a paid IP transit provider that delivers traffic on behalf of another network.
Peering is the gold standard: one hop between networks, meaning lower latency and higher reliability. Transit is a necessary fallback for reaching the “long tail” of networks—there are over 100,000 Autonomous Systems on the Internet, and no network connects directly to all of them.
Cloudflare, for example, peers with more than 12,000 networks settlement-free, and connects at 287 IXPs worldwide—the second most of any network, according to Hurricane Electric. But many of Europe’s largest incumbent telcos decline to peer settlement-free, forcing traffic onto transit links. That works adequately most of the time, but it isn’t optimal.
The regulatory question is whether this market needs fixing. Europe’s transit market is competitive: according to Telegeography, the lowest prices for 100 GigE IP transit in Europe were $0.06 per Mbps per month—consistent with US pricing and a sign of a healthy ecosystem. Low wholesale bandwidth prices lower entry barriers for small ISPs and keep costs down for all types of applications and services. Europe’s own telecom regulator, BEREC, has concluded that the Internet “has proven its ability to cope with increasing traffic volumes… without a need for regulatory intervention.”
The incumbent telcos seeking those fees don’t see it that way. In practice, large incumbents often demand paid peering fees that are multiples of what a CAP would pay a transit network for equivalent service. At those prices, many networks—including Cloudflare—route around them via transit. The telcos are now seeking regulation to force CAPs into paid peering at artificially high prices.
That would likely set a de facto price for all interconnection. If telcos can extract above-market rates from the largest CAPs, there is little reason for them to accept lower rates from smaller networks or transits. Wholesale prices would stop falling, and the increases would be passed on to small businesses and consumers.
Who gets the fast lane?
A network usage fee regime would make the consumer experience worse, not better. Even under existing net neutrality rules, some large telcos already treat transit traffic as second-class. In November 2022, the Hungarian tech site HWSW documented recurring congestion between Deutsche Telekom and its transit providers, affecting users of Magyar Telekom and other DT subsidiaries:
Network problem that exists during the fairly well-defined period, mostly between 4 p.m. and midnight Hungarian time, … due to congestion in the connection (Level3) between Deutsche Telekom, the parent company that operates Magyar Telekom's international peering routes, and Cloudflare, therefore it does not only affect Hungarian subscribers, but occurs to a greater or lesser extent at all DT subsidiaries that, like Magyar Telekom, are linked to the parent company. (translated by Google Translate)
This isn’t an isolated incident. In 2015, transit provider Cogent sued Deutsche Telekom, accusing it of “refusing to increase the capacity of the interconnection ports that allow the exchange of traffic.”
If the biggest CAPs pay for interconnection, their traffic will take priority—and all other traffic will be relegated to the congested links that remain. The perverse result is that regulation aimed at extracting fees from Big Tech would entrench Big Tech by giving its customers a better experience at everyone else’s expense.
There’s also a legal problem. Article 3(3) of the EU’s Open Internet Regulation requires internet access providers to treat all traffic equally “irrespective of the sender and receiver, the content accessed or distributed, the applications or services used or provided.” Fees from certain content sources in exchange for private interconnection paths look like a plain-language breach of that provision.
The quiet engine of the Internet: settlement-free peering
Most of the Internet’s interconnection happens settlement-free. A survey of over 10,000 ISPs found that 99.99% of exchanged traffic moves on settlement-free terms. Networks peer directly because it benefits both sides: lower latency for users, more resilience through multiple paths, and local traffic exchange that keeps data from aggregating at regional hubs.
The quality of a connection depends more on latency than raw bandwidth. Delivering content from a local point of presence avoids the latency penalty of hauling data across a country—or to a regional hub and back. Yet in parts of Europe, traffic is routed through transit providers at central Internet hubs rather than exchanged locally.
Germany illustrates the problem. Traffic that could be exchanged at one of Cloudflare’s five German points of presence instead flows through Frankfurt. A consumer in the northeast of the country experiences almost double the latency of someone closer to Frankfurt, and their traffic is exposed to congestion on a small number of transit links. Settlement-free peering at all five locations would give every German consumer a comparable experience.
The issue isn’t network capacity—fibre backbones aren’t constrained, and ports can be added cheaply—but where interconnection happens. If a telco carries subscribers’ requests to a regional hub to exchange traffic with a video provider, the data must traverse the telco’s own backbone twice. Local interconnection or embedded caches, which most CDNs including Cloudflare offer, eliminate that backhaul. Embedding a cache means installing the CDN’s hardware at aggregation points inside the telco’s network; subscriber requests are then answered locally and never reach a regional hub. Cloudflare has hundreds of such deployments with telcos globally.
What’s at stake for Europe’s digital future
Despite the hesitancy of large incumbents to peer settlement-free, the interconnection market is working. Competitive transit pricing is a feature that benefits European consumers and small businesses. Forcing CAPs into paid peering agreements would not meet the Commission’s Digital Decade goals; it would degrade the Internet experience for the very constituencies those goals are meant to serve.
Concern isn’t limited to the companies that would pay the fees. Digital rights groups, the Internet Society, European video-on-demand providers and broadcasters, Internet exchanges, mobile operators, several European governments, and Members of the European Parliament have all spoken out against interconnection regulation.
The Commission’s consultation remains open for submissions until 19 May. Anyone—not just telecom lawyers—can submit a narrative response. The right message: ubiquitous Gigabit connectivity is a worthy objective, but the route the Commission is considering leads in the opposite direction.
