Another Round of Net Neutrality

The Bush administration has fired its latest salvo on net neutrality, with an in-depth report by the FTC (actual report here). Their conclusion: let's wait and see. It's a 170 page report and covers a lot of ground. I can hardly respond to everything in it, but I would like to make a few comments.

First, the FTC does a solid job of covering the arguments in favor of net neutrality (pages 56-64). They didn't make any effort to soft-peddle the substantive arguments or build the strawman positions many net neutrality critics are so fond of. Most significantly they covered the concern that prioritization of ISP-provided content and other specially licensed preferred content would tend to recreate the walled gardens of the pre-Internet America Online and Prodigy days. They discussed the problem that lack of competition in the last mile makes the market an unreliable regulator of ISP behavior. Also they noted the argument that increasing bandwidth (say to 100 Mb/s) could make the issue of congestion largely disappear (and concerns over net neutrality with it).

The report also summarizes arguments against net neutraity (pages 64-69). Most of them are, I think, pretty weak. They raise the usual point about the necessity of non-neutrality to deal with small numbers of users sucking down large amounts of network resources for filesharing and the like. This really has little to do with net neutrality. Net neutrality is about ISP's throttling the content end of the connection, not the user end. Even under most proposed neutrality regimes ISP's would be free to throttle users (within the terms of their service agreements) when users abuse the network. They also raise the point that different types of data (web pages versus VOIP versus streaming video) may need different service levels. This again, as I pointed out in my last net neutrality post, is not prohibited by net neutrality. It's ok to give VOIP traffic higher priority than web traffic, as long as no particular VOIP provider is preferenced over another.

The net neutrality opponents argue that net neutrality is necessary to allow service providers to capture enough revenue to fund the buildout of faster networks needed for advanced services. This seems facially plausible, but breaks down on further analysis. First, if what the ISP's want is a metering system, so that the users of advanced services bear more of the cost, there is nothing to stop them from simply offering those capabilities directly to end users for a price. Already some providers offer multiple tiers of internet service. Additionally, the net neutrality model doesn't appear to be one well suited to driving growth in network capacity. Non-neutrality revenue is built on bandwidth scarcity. If bandwidth isn't scarce, no content provider pays an ISP for preferred service. I find it hard to believe that if ISP's start to have significant new revenue streams coming in predicated on the scarcity of their network resources, they would turn around and invest that income in expanding their network resources and thereby undercut their new revenue model. In a non-neutral market it pays to under-supply bandwidth.

Finally, perhaps the most critical topic of discussion is the state of competition in broadband internet service (starts on page 98). If internet service was truly a competitive market, there would be no need for net neutrality regulation, as consumer demand could sufficiently regulate the market to prevent abuse by ISP's.

The report starts off by noting how much broadband prices have fallen and how much speed has increased in the last six or seven years. Certainly it's true that this has happened, but it needs to be put into perspective. This is, of course, a business built on technology that is improving at remarkable rates. Look at how much PC's have advanced in the past decade, how much faster the processors are, how much more RAM they have, how much more harddrive capacity, how much better video processing, and how much prices have fallen. By comparison, the rate of change in broadband internet service is glacial. Or you could compare cost and speed improvements in broadband service in the US with broadband services in Japan or Korea or France, where, on a price/performance basis, we've been left in the dust. Yes, the market is improving, but it would be an appalling failure if it didn't, and the rate at which it is improving does not bode well for the state of competition in the broadband market.

The other key point on competition is the matter of new entrants to cover the "last mile". Opponents of net neutrality argue that the cable/phone company duopoly is being challenge by other service providers using satellite, broadband over powerlines (BPL), or wireless systems. Each of these technologies, unfortunately, faces crucial barriers. Satellite is cursed with poor latency. As fast as the speed of light is, bouncing a signal off a satellite down to a receiving station, then back up to the satellite and back down to the user takes time, enough time to make satellite unusable for most interactive advanced services. BPL simply doesn't appear to be making the cut in terms of cost of deployment and performance level to entice many power companies to want to try rolling it out. It simply isn't going to have the cost/performance ratio to make it competitive with cable or DSL any time in the near future (if ever). For wireless the challenge is spectrum. Wireless will be highly competitive in the market for internet service at speeds of 10 Mb/s or less. But it does not appear that there is enough spectrum for widespread use at speeds much greater than that. As people become more accustomed to high speed cable/DSL/fiber service they will find wireless insufficient for standard residential use (and for most uses that implicate net neutrality concerns). Cable and DSL can provide speeds in the 30 Mb/s range (although no US providers offer this at present), and fiber will go up to 100 Mb/s. I'm afraid we are going to be stuck with a duopoly for some time. The market is not going to bail us out of this mess (although it could if federal regulators forced the cable and phone companies to open up their last mile networks to other competitors).