On Network Neutrality

The Wall Street Journal sparked new interest in the issue of network neutrality when it published an article claiming Google, Yahoo!, Microsoft and others have quietly removed their support for net neutrality. They also claimed that famed techno-political activist Lawrence Lessig had changed his stance; he, however, says his position has always been as they described, and it's not changed recently. So I feel I should lay out my position.

The basic question in network neutrality is: should content providers be able to pay more to guarantee faster delivery of their content? For instance, Google's plan is to pay AT&T, Comcast and others to put servers actually inside their networks, thus delivering content faster. This sounds pretty innocuous, right? The market will solve, and better content providers will have the money to pay for better service. Everybody wins!

Pay for play

Except that the flip side of the question is: should Internet access providers be able to charge content providers for delivery of their content? As soon as you open the door to say that Comcast can charge Google $5/month to get priority delivery, you also give Comcast the ability to approach Microsoft, Yahoo! and everybody else and say "you're now getting 'standard service', which we've just redefined as 'painfully slow'. If you want 'super premium' service, which feels a lot like standard service used to feel, you need to pay us $5/month".

Would that ever happen? I'm sure no Comcast customer would dispute it; the company is not known for its great service. Additionally, the company has a history of lying about filtering their customer's Internet access, so we're not dealing with Mr. Ethical here. Furthermore, within any given municipality, there are usually only two or three viable Internet access providers -- often just one. So we can't rely on competition to keep these guys honest. It's a clear case for regulation: having dozens of companies compete to lay wire to people's houses is impractical, so instead you have to regulate the behaviour of the ones that you do.

It's a series of tubes

In short, allowing Internet providers to charge for access makes the incentives work the wrong way: instead of trying to reward their customers with a great Internet experience, it becomes in their interest to give customers a crappy default Internet experience to encourage providers to pay. Newcomers will have a crappy experience by default. So if the next Google comes along with a radically better search experience, it won't matter, because Google can copy them, and because Google has a ton of money to pay Internet providers, they can make their product seem better even if it's not.

The stifling effect this would have on competition would be undesirable, to say the least, and might even put the US at a competitive disadvantage: the US already significantly lags other developed nations in Internet access speed and penetration. This is no slight difference -- Japan's average internet access speed is more than 30 times faster than in the United States. These ISPs are doing a crappy job, and do not need any further rewards.

Lessig's analogy for allowing ISPs to charge for faster service, quoted in the WSJ article, is:

There are good reasons to be able to prioritize traffic. If everyone had to pay the same rates for postal service, than you wouldn't be able to differentiate between sending a greeting card to your grandma versus sending an overnight letter to your lawyer.

But I don't think this is an accurate analogy. To paraphrase Ted Stevens, the Internet is not like a mail truck. It's a series of tubes -- or even better, it's like roads. Like Internet pipes, it's not practical to build separate roads to everybody's house. They're a common resource. And while you can pay for your stuff to be delivered faster, it's travelling on the same roads and obeying the same laws as everybody else -- the guys charging you for overnight delivery are just arranging things better, consolidating delivery and logistics. You could think of this as faster servers, and better compression methods. But UPS does not get its own lane on the highway, no matter how much it pays.

What about air freight? Not everything has to travel by road. Surely the analogy breaks down? I'm not sure it does. Air freight is sort of like building your own trans-pacific cable to carry your Internet traffic. It's crazy fast, crazy expensive, and doesn't crowd anybody else out. And like air freight, the last bit before it gets to your house has to go over the same old roads.

No closing the door behind you

As I've said before, Internet access is not just a technology, it's a form of speech, like publishing, and it will be as fundamental to the next 100 years as publishing has been in the previous hundred. It needs to be encouraged and protected from vested interests who would stifle that speech. The big Internet players of today will not be the big players of tomorrow, and that's the way it should be. We should not provide an economic structure that will allow them to cling to supremacy any longer than they deserve to.