Born Weird

I’ve been meaning to post for a while now on this subject, but haven’t had the time.

And frankly, I still don’t. But given the furor over the FCC’s notice of proposed rulemaking today, I feel the need to quickly dash off a some thoughts. Hopefully at some point I’ll be able to do a more thorough treatise with actual citations and numbers, but this will have to suffice for now. Here goes.

What is Net Neutrality?

Many ardent supporters and fervent opponents of Net Neutrality seem to be rather confused about what exactly Net Neutrality is.

At its essence, net neutrality is the principle that a network ought to treat data packets equally and without preferential treatment. How far that “neutrality” should extend is open to debate even among NN advocates. At the very minimum, most agree that to be considered neutral, a network must not discriminate on the basis of “source, ownership or destination.” This, I believe, is the most sensible definition, and the one that most closely resembles current policy proposals on the table.

Some die-hard NN supporters take it further and argue that discrimination based on data type should also be verboten, which is to say, no allowances can be made by the network operator with regard to differing quality of service requirements or expectations; “first come, first served” would apply across the board. Personally I think this is a bad idea, but that’s beside the point.

What Net Neutrality Isn’t

There’s a lot than can be said here. But in brief, the flavor of NN currently under debate isn’t:

1. A particularly new idea — it’s part of a long regulatory history.
2. A “Fairness Doctrine” for the Internet — it’s content neutral, not content balance-enforcing.
3. A goverment takeover of the Internet’s infrastructure — no transfer of ownership of any kind is involved.
4. A ban on “reasonable network management” practices — it still allows for QoS tiering, bandwidth caps, etc.
5. A ban on consumption-based pricing — nothing about NN prohibits “pay as you go” pricing.
6. A substitute for true competition.
7. A government diktat regarding broadband capacity.

Real World Examples

To move out of the abstract here, let’s think about what would (and wouldn’t) constitute a NN violation. If an ISP (e.g. Comcast, Verizon, AT&T, Time Warner, whoever provides your “always on” Internet Service) were to block you from accessing a particular website, that would be a violation (discrimination based on source). If your ISP were to deliberate degrade the quality of Vonage VOIP calls over its network, that’d be a violation (discrimination based on ownership). Or if they were to fake “reset packets” to prevent you from using BitTorrent. Or if they capped the bitrate on streaming video from Instant Netflix, but not from Amazon On Demand.

It wouldn’t be a violation if your ISP slowed down text data packets from all websites you’re surfing in order to guarantee speedy delivering of all the video being delivered to your home — that’s QoS prioritization, and that’s okay provided all video packets are being treated equally to one another, all text packets are being equally to one another, and so on. It also wouldn’t be a violation if your ISP started charging you extra for exceeding a given amount of data transfer in a month (provided all data counts equally toward that cap), or slowing down your connection after you exceed a quota; those “across the board” network management techniques are clearly reasonable and nondiscriminatory.

Net Neutrality in History

While the term “net neutrality” is fairly new, the idea behind it — that a network owner can’t discriminate for or against content traveling over that network — isn’t new at all. It goes back to telephone service, which in turn goes back to the telegraph, which in turn goes back to railroads, which in turn goes back to other transportation networks and English common law. In each of these cases, network owners have been treated by the government as “common carriers,” thus subjecting them to regulations regarding nondiscrimination, interconnection, and a host of other things.

These regulations historically have been in place to prevent all sorts of various misbehavior on the part of network owners, such as AT&T wanting to charge a different rate (or ban altogether) phone calls used to transmit a fax or using modems, or a railroad wanting to strike deals to carry produce from certain suppliers quickly while denying other farmers access.

The reason we’re even having this debate now is due to a long sordid regulatory history of the FCC which I’ll summarize here very casually. Basically, before there was Internet, there were just phone lines and cable TV lines. Each had their own regulatory classification and thus they were subject to different regulations, with phone service clearly subject to common carrier rules and cable TV not. When phone companies and cable companies starting offering Internet service, they started competing with one another and asked for regulatory parity. In a court case known as Brand X, it was decided that both phone and cable Internet would be treated more like cable service and less like phone service (I’m glossing over a lot of details here).

So where once the principles of Net Neutrality were essentially in place (at least for DSL), they are no longer. Of course, whether or not they were in place earlier isn’t so much of an issue, because there was no good reason to discriminate not that long ago; the marketplace was very different then. So what’s changed? Increased competition, and decreased competition.

Why We Need NN #1: Increased Application Competition Provides Incentive to Discriminate

One reason the big ISPs are so opposed to Net Neutrality is that they’re no longer just providing an empty pipe for Internet use: they’re also providing their own services (aka applications) on top of that Internet connection. Comcast, for example, offers Digital Voice, a Internet-based phone service. AT&T offers U-verse, which provides TV service using the same protocols that underly the Internet.

And traditional Internet content providers are offering fare that directly competes with the traditional bundled offerings of today’s ISPs: Vonage and Skype offer phone service, Netflix and Amazon offer video-on-demand, etc.

These competing offerings provide a financial incentive for network providers to discriminate — why allow your competitors to use your infrastructure to compete with you? Of course, you don’t have to block a competitor to gain a competitive advantage; you can tinker with their application’s performance  just enough to make yours more attractive, avoiding the PR disaster affiliated with being a “censor” while still essentially picking winners and losers.

Why We Need NN #2: Decreased ISP Competition Removes Penalty for Discriminating

Competitive markets are great things: when players misbehave, they are punished by the marketplace. Ideally, rather than stepping in with government regulation to guarantee nondiscrimination, we could rely on the marketplace, with consumers switching providers whenever their ISP starting acting in a discriminatory manner that wasn’t in consumers’ best interest — for example, BitTorrent users could all switch from Comcast to Verizon when Comcast banned BitTorrent.

Leaving aside the fact that it can be very difficult for a consumer at the end point of a network to know why an application is performing badly (it may not be due to discrimination after all), the fact remains that the local broadband market in the United States isn’t particularly competitive. I don’t have the numbers handy, but almost no one in the country has a choice of more than two broadband providers: their local cable company, and their local phone company. That’s it; broadband over satellite has yet to emerge as a viable competitor/substitute for wireline service. And millions of Americans don’t even have a choice of two.

While some people may consider a choice between two companies robust competition, a pair of players is generally considered a duopoly and leaves consumers with very little recourse when, say, discrimination has run amok. If both Comcast and Verizon were to block Vonage, I would be completely unable to remain a Vonage customer. Competition is clearly not sufficiently vibrant to leave nondiscrimination enforcement to natural market behaviors.

Earlier, there was more competition in the ISP market due to an FCC requirement of “Unbundled Network Elements,” which essentially allowed competitive ISPs to offer service by piggybacking on phone companies’ DSL infrastructure. You may even remember when you had a choice of several DSL providers (e.g. Earthlink) — but then the courts struck down unbundling, and DSL competition evaporated.

The Capacity Conundrum

Some of those opposed to Net Neutrality have suggested that NN will have an adverse impact on broadband capacity. I’m not sure what the logic behind that argument is. I guess the thinking is that Net Neutrality will reduce the amount of rent-seeking (aka “innovation”) on the part of ISPs who want to develop tiered pricing structures that allow for higher profit per unit of data transferred, and that this reduced profitability will have a suppressive effect on high-capacity rollout.

There’s little evidence to support that, however. For one thing, Verizon is busy rolling out FiOS, and AT&T U-verse, in an environment where discrimination is being held at bay by the current regulatory uncertainty (it’s easier to argue that NN isn’t necessary if you’re not busy violating it, so companies are largely holding off until a decision is made one way or the other), and cable providers are rolling out the high-capacity DOCSIS 3.0 (which doesn’t involve laying new last mile wires).

For another, because Net Neutrality will necessarily put an end to more sophisticated (or insidious) traffic shaping, the only/best way to ensure quality service will be to fatten the pipe (rather than manipulate how things travel through it). So in that sense, NN will actually incentivize increased capacity.

The Dark Side of Net Neutrality

I’m willing to admit that Net Neutrality is an imperfect solution to the problem of subtle (and obvious) discrimination. It’s a poor substitute for real competition. The law of unintended consequences has a very long arm in tech policy. And if anyone is capable of screwing things up, it’s the federal government. There are certainly plenty of opportunities to do Net Neutrality wrong, from being too heavy-handed, to having too light of a touch, to being just too vague for network operators to know what they can and can’t do. Missteps could have a suppressive effect on online freedom or broadband deployment.

But the alternative — doing nothing — is far worse, as once the ISPs get entrenched in new revenue models that rely on discrimination, there will be no undoing them, and the online world will have more gatekeepers than ever before. If you think the left-wing public interest advocates that surround the FCC are powerful enough to dismantle the business model of the multibillion-dollar telecommunications companies, you’re a fool.

The good news is that the FCC doesn’t work like Congress — there are no Patriot Act / health care reform bills that get pushed through before anyone has a chance to read them. The FCC is a very deliberative body, and if anything is even more transparent now than it has been in the past (which admittedly isn’t saying all that much). Each step of the process has a public comment period and anything the FCC does is likely to be challenged in court. Even if there are some bad intentions on the part of some advocates on the left — always a possibility — nothing insidious is going to stick.