Intelligent networks are better for congestion than pricing
Adam Thierer has written his take on Farhad Manjoo’s article “The iPhone is not an all-you-can-eat buffet” where the two agree that congestion pricing is a great way to control congestion on 3G networks. Manjoo argues that most smart phone users who are only using 40 to 80 megabytes per month are being crowded out by iPhone users who are averaging 400 megabytes per month. While I agree that usage capped pricing is one solution to the problem, I think there may be better alternatives that will solve the problem in a more surgical manner and be more popular with consumers.
The fact of the matter is that metered pricing or usage capped pricing are very unpopular in America and the U.S. market is one of the few countries in the world that has largely rejected usage capped pricing. While I’ve always thought that usage caps were the least desirable mechanism to deal with congestion, I can’t deny that it has the benefit of reducing prices for a large majority of subscribers and I certainly don’t want to see the business model outlawed by regulators. But since this discussion is about the 3G wireless congestion problem, I think it’s worth discussing alternatives especially if they are better at handling congestion and if they are more popular with consumers.
The better solution to congestion on networks (especially wireless networks) is to prioritize low usage subscribers over high usage subscribers and this is what I described as a form of “reasonable discrimination” that regulators must permit. This is exactly how the new Comcast “fair share” network management system works and it was approved by the FCC to succeed the controversial TCP reset system. More importantly, prioritizing traffic is more precise at managing instantaneous congestion whereas pricing only manages average congestion.
The premise of the system is that networks are shared and everyone competes for network capacity while paying the same for the service yet some consumers take 10 times the network resources. The fair thing to do is allow the consumers who use the least capacity to have priority over consumers who consume the most so that everyone has a fair shot at getting something from the network. The high usage consumers already had their turn so to speak, and they’ll continue to get their turn but they’ll get shorter turns than the customers who hadn’t had any turn at all. Even under this type of system, the heavier users will still get a larger share of the network while keeping their generous usage allowances but the majority of other users who pay more of the bill but use less of the network will get a fair shot at the system when they need it.
This is not to say pricing has no role and it could be a part of a prioritized network. Those who wish to have larger priority allowances and larger usage caps can still be offered premium services at a premium rate. That means they can be heavy users without having to cede priority to the lower usage subscribers. If there is sufficient demand for the premium service, it would justify more infrastructure.
We don’t need to be network engineers to intuitively know that this is a fundamentally fair system. While the system discriminates against heavy users, it is fair discrimination needed to make the system fairer. The problem is that there are regulators who are trying to outlaw any kind of discrimination under the guise of “Network Neutrality” even if the discrimination is fair and this would be a tragic mistake that would leave usage based pricing as the only mechanism available. The market has shown that it knows how to pick the best business model and neither congestion pricing or network prioritization should be prohibited by regulations.