Although building more bandwidth is always the preferred solution, it seems clear that traffic-shaping has a place in network optimization. I'm even open to the idea that the market should help determine how it's implemented.

The problem is that it can be very difficult to differentiate between policies that will increase the overall efficiency of the internet and those that will simply increase the economic efficiency of the network operators. If charging a premium for expedited delivery of VoIP traffic will allow an ISP to build out more capacity, great — it's easy to see how this arrangement can result in an a net increase in wealth. The pie gets bigger, as Matt likes to say.

On the other hand, if they choose to deliver these VoIP packets more quickly by throttling their other customers' web traffic, well, that's not as positive a development. It could be the best possible decision, but it's impossible to say without a hard look at the facts. That's where network engineers come in — and where the profit motive would complicate the decision in the absence of network neutrality. That's because it'd be easy to use network optimization as a cover for rent-seeking behavior. You'd be crazy to think that, given the chance, the ISPs wouldn't try to exploit such a situation. Comcast's arbitrary bandwidth caps for its customers are already implemented in an appalling manner. The business of network optimization is not one that's conducted with an enormous amount of scruples when profit becomes intertwined with more mundane considerations of jitter, throughput and latency.

And that, to me, is the real difficulty in this debate: how do we encourage the competitive network practices that will benefit us all, while preventing the ones that would allow network-owners to enrich themselves without providing benefits to consumers? There are a few things that I would like to see happen to encourage the former and make the latter more difficult.

First, consumer bandwidth should be metered after a prepaid cap. The ISPs should be in the business of selling us internet use. Right now their preferred scenario is for users to sign up for broadband and then use it to check their email monthly (and hopefully never call tech support). They should want us to use their networks — that will spur them to build better ones.

Second, admission to different classes of service should be done on the basis of the nature of the service — not the identity of the business buying the service. If Comcast tried to push Skype and Vonage into a bidding war to see whose website would remain usably-speedy, it'd be outrageous. If they specified a rate for prioritized VoIP service and allowed anyone to purchase it — well, that would make a lot more sense.

Relatedly, it'd probably be a good idea to start thinking about antitrust standards for ISPs, particularly as network and entertainment service becomes increasingly merged. I'm far from an expert on these issues, but I do know that they're going to become more and more relevant. Imagine if Time Warner Cable selectively sped up the downloads of Warner Bros. movies from online movie stores. Or if it allowed only its own movies to be stored on the caching servers at their cablemodem central offices. Or if TW simply degraded the service to all movie stores except the one run by their preferred corporate partner. Just a business leveraging its (ugh) synergies? Maybe. But it'd also be a good way to freeze competitors and new entrants out of the market, utilizing a business advantage not available to anyone else.

Third and finally, legislative roadblocks to municipal wifi should be thrown out. I have no idea if publicly-run broadband is the model that will ultimately make the most sense for consumers and communities. But it seems obvious to me that internet service is more like a utility than anything else, and that the market for it ought to have the option, at least, of moving into a structure like the one used for those services. More importantly, the more choices consumers have for internet access, the more competitive the ISP market will become, reducing the need for regulation to keep our current pair of broadband options free and usable.

Even in the headiest days of utility deregulation we didn't let the water company lower the pressure in YMCA showers just because Gold's Gym had signed up for premium service. I doubt we'll ever see a similarly egregious abuse by the network operators, but there's no reason to open ourselves up to one.