Net Neutrality: This Time it’s Different

I read a blog post today that argued the net neutrality fight is one long, uniform battle that’s been going on since 1998, when it was supposedly called “open access”. That sentiment says a lot because it’s just as sensible to say each regulatory debate over net neutrality is very different from all the others. While some of the arguments are always recycled, each debate has to digest some new facts. So I think it’s useful to take a look at what’s different this time around.

This Time is Different

We used to understand net neutrality to mean that ISPs have to treat all packets of information the same way: Whatever comes in goes out, in the same order as every other packet, more or less. This idea was modified by the notion that it might be considered “reasonable network management” for an ISP to react to temporary congestion differently. And there has also been some belief that it might be reasonable for ISPs to treat voice packets different from web packets, but the FCC has never been clear about that.

This idea of fairness is expressed in the three bright-line bans on blocking, throttling, and prioritization based on the source, destination, or application associated with each packet. But the current FCC regulations take two large steps. First, they place the entire broadband industry under Title II, creating a huge gulf between ISPs and the rest of the Internet. And second, the last FCC created a “general conduct standard” that allows the agency to take action against ISPs in the absence of published regulations.

Title II and the General Conduct Standard

So this time around we’re not arguing bright-line rules, we’re arguing about Title II and the amorphous grant of authority the FCC gave itself. Title II enabled the FCC to create an unlevel playing field in the advertising market, under the guise of consumer privacy. This action prompted Congress to actually strike down an FCC regulation under the Congressional Review Act, something that never happens.

The general conduct standard calls the entire enterprise of ex ante regulation via notice and comment rulemaking into question. There’s very little guidance about disfavored conduct in the 2015 Open Internet Order. If the FCC is right that some forms of bad behavior are unknowable in advance but evident when they occur, it undermines its own role. If the FCC’s reasoning for the general conduct is correct, then it follows that net neutrality enforcement should shift to the FTC, an agency that does not use the notice-and-comment process.

Neither Title II nor general conduct were debated in 2014 because then-chairman Tom Wheeler didn’t fill us in on where he was going until he got there. So this is a new discussion of those two questions.

We Now Know that Net Neutrality is Unenforceable

As much as we all may love the idea that ISPs should not block, throttle, or prioritize, it’s now evident the FCC has no way of knowing whether these activities are taking place. This is thanks to a study done by Neil Davies, the British computer scientist, for Ofcom, the UK’s FCC, in 2015. We discussed this study in a podcast with Dr. Davies in 2015.

In technical terms, net neutrality bans “differential traffic management”. But Davies declares “we must conclude that there is no tool or combination of tools currently available that is suitable for practical use” in this endeavor. This is because measurements can only be taken at individual endpoints and the Internet’s behavior is so variable that it’s impossible to distinguish signals of differential treatment of traffic from statistical noise.

Given that net neutrality violations are fundamentally undetectable, questions about legal underpinnings for regulation take on an entirely different character. Instead of figuring out how to enforce a rule, we need to focus more on what we mean by net neutrality violations and how to find them.

If net neutrality violations are undetectable today, it follows that they have never really been detectable. The only clear case of differential treatment that we’re certain about is the case of Comcast reducing the number of TCP streams that Bittorrent could use for unattended seeding in 2008, and even that was merely a temporary stopgap to prevent Internet piracy from slamming Vonage calls.

We can certainly describe net neutrality in simple terms, but that doesn’t mean our consensus descriptions are objectively meaningful. This is a huge problem for net neutrality supporters, and one that is largely unacknowledged.

ISPs aren’t Credible Threats to Internet Freedom

The entire Internet ecosystem has changed dramatically since Tim Wu presented his first net neutrality paper in 2003 and FCC Chairman Michael Powell articulated his Four Internet Freedoms in 2004. Net neutrality assumes that the Internet is open and free only as long as ISPs don’t unfairly interfere with its operation. In 2003, ISPs were the biggest players in the ecosystem and the Slick Six (Amazon, Apple, Facebook, Google, Netflix, and Microsoft) were, apart from Microsoft, lovable rebels.

Despite the mergers and acquisitions that have made ISPs larger, they have less influence over the Internet than ever before. The argument that any ISP can actually block access to the Slick Six without suffering major marketplace harm is simply laughable today. Are the startups of today worried about ISPs slowing down their packets or about one of the mega-giants reducing their entire value proposition to a mere feature?

In fact, the the vertical mergers between ISPs and content creators suggest that ISPs are facing competition for basic access services in the near future. It’s hardly a done deal, but if 5G comes anywhere close to expectations consumers are going to have 3 to 5 choices for basic Internet in most markets. That development by itself suggests that the basic premise of net neutrality – perpetual lack of ISP competition – is unsound.

The Internet has Different Issues

Apart from the consolidation of edge services, the detectability of net neutrality violations, and the legal questions, net neutrality has to take a back seat the real technical and financial issues facing the Internet. While we used to worry about deployment and download speeds, these problems have become academic in most markets.

FCC data says that 98% of developed Census blocks have access to two or more broadband providers offering speeds of 10 Mbps or more. We’re concerned about ensuring that rural areas have access to high-speed, fixed location services and that we will have the infrastructure in place for 5G in time for speedy deployment. Markets are generally competitive regardless of the Wheeler FCC’s cooking the books by defining broadband at an arbitrary 25 Mbps in order to erase competition.

We’re also concerned about the fact that Google and Facebook are eating up the market for new Internet ads, because that compromises the ability of new ad-funded services to thrive. We’re concerned that there are so many trackers embedded in web pages that pages are slow to load and hard to navigate.

And we’re concerned about the fact that the Internet is still unsafe to use in many respects, too intrusive with respect to privacy, and stagnant in terms of new applications. These issues don’t need to be explained to Internet users the way net neutrality does: they’re right in our faces.

So the sooner we can tie a bow on net neutrality the sooner we can get back to solving real problems and addressing real issues.