Well, the FCC is back to the net neutrality drawing board, drawing inevitable comparisons with the classic movie. The link I referenced is a good primer on the concept, written by someone who, like me, is skeptical about the concept. It’s not that I don’t believe we need some regulation, but it’s important that the rulemaking not go too far. Here is a link to the current fact sheet (April 4, 2024) on the activity, and it suggests that the worst risks might not be realized. But the new rule still creates issues that I doubt the FCC will be able to resolve. In addition, if there is an administration change, we can be sure we’ll end up with a replacement rule, and that raises a point I’ll get to below.
I think we can divide neutrality goals into three broad categories. First, there are rules to prevent ISPs from throttling traffic for selfish or opportunistic reasons, which is the issue my first reference link talks about. Second, there are rules that relate to how ISPs settle for peering traffic, and third there are rules that aim to build a fence around one of the first two issues to ensure that an ISP doesn’t dodge the effect of a rule with a clever service strategy.
I have to disagree that Comcast’s decision to throttle peer sharing protocols to manage the limitations of upload traffic on their cable infrastructure was the dawn of the neutrality debate, while I do agree that it gave a name to it. The real start of the neutrality problem was the launching of the worldwide web, the concept of web severs and browsers that started with Mosiac, twelve years before the Comcast kerfuffle. The Web made the Internet about experience delivery, which is inherently asymmetrical in traffic terms…and in business terms. The networking needed to connect users is way more expensive than the one needed to connect content, and since the Internet was launched in a bill-and-keep model, that meant that “retail ISPs” were certain to be under business pressure eventually. The question of settlement among ISPs came up within a few years of the Web’s launch, and the solution that came along was peering agreements among ISPs. Peering agreements are the only permitted source of settlement among ISPs and with content providers.
What the FCC has done is first to restore the previous classification of the Internet as a telecommunications service, primarily to give the FCC regulatory authority. In a perfect world, what should happen is a new telecommunications act, but obviously getting that done in the current political climate would be very difficult. The disputes over this point are, IMHO, primarily from parties who don’t want the regulations that follow. The FCC also takes steps to forebear applying many of the rules (like tariffs) to ISPs that are applied to other telecommunications providers.
Those regulations start with “straightforward, clear rules that prohibit blocking, throttling, or engaging in paid or affiliated prioritization arrangements”, and here we see a mingling of the first of my three rule classes with the third. The original goal of preventing traffic discrimination could be circumvented by introducing payment for better treatment, which when the rest of traffic is “best efforts” has the effect of implicit throttling. You can easily see that big players could pay for delivery of their stuff where startups might find this a burden.
The risk posed here is that there are applications that cannot be reliably supported on best-efforts service. To forbid paid prioritization is to foreclose these applications from using the Internet. Given that the Internet is declared to be a “telecommunications service”, would it not make sense that it be considered the telecommunications service, discouraging alternate services that could only create a form of overbuild. And what happens if such an alternate service proposes to carry what is now traditional Internet traffic, given that it was launched to provide better QoS?
You can also argue that when an Internet user buys gigabit broadband rather than 100 Mb, they are engaging in paid prioritization. Is it inconsistent to allow them to do that, but say that paying for a “fast lane” or specific QoS is potentially anti-competitive. In early discussions on this topic, it was suggested that the rule allow users to pay for a fast lane, but not content providers, on the likely correct theory that it was this third-party payment that posed the real risk. Then, of course, you have to prevent content players from reimbursing their users out of their own fees.
The proposed order does recognize that “edge providers” exist and may need special services beyond basic Internet access service (BIAS, as the FCC terms it). This portion seems to at least attempt to address IoT/M2M applications and perhaps edge computing as well. The FCC seems to aim at allowing at least a “special lane” for edge traffic, but says that what creates the difference is the user of the service, which would suggest that if you used an “edge provider” service to deliver something the Internet would customarily handle, you are now serving an Internet user, thus making it a BIAS, and you lose your ability to offer special handling.
We now come to the ever popular topic of Internet peering. There have been proposals over time that include everything from having no peering charges at all (open free peering in other words) to have a tariff peering and even traffic oriented settlement made mandatory. The FCC seems to be taking a position that the current commercial peering process, where peering agreements, including charges, are up to the companies themselves is the preferred strategy. This is certainly the least markets disruptive approach that could be taken, but it obviously leaves questions open regarding the relative market power and health of the ISPs that support retail customers versus those that support content providers. IMHO, the FCC peering stance suggests at least that they would not favor big-tech subsidies of the type EU telcos have sought.
What this order would do, as drafted, seems to be returning Internet regulation to a prior state, restoring “neutrality”. However, most ISPs and tech companies tell me that they didn’t really see much difference in the “non-neutral” period and the “neutral” one, and even the FCC notes that state neutrality rules and the chances that rules would change every time the party in power changed, have dampened any desire to specialize Internet services, shifting them from the long-term practices that have emerged. To me, this means that any fundamental policy shift can’t be imposed by the FCC, but will require another Telecom Act to amend or replace what we got in 1996.
