DC Circuit court rules in Comcast case, leaves the FCC a job to do

Today's ruling in Comcast v. FCC will certainly change the terms of debate over network neutrality, but the win for Comcast is not as far-reaching as headlines make it appear. The DC Circuit court didn't say, "You folks at the Federal Communications Commission have no right to tell any Internet provider what to do without Congressional approval." It said, rather, "You folks at the FCC didn't make good arguments to prove that your rights extend to stopping Comcast's particular behavior."

I am not a lawyer, but to say what happens next will take less of a lawyer than a fortune-teller. I wouldn't presume to say whether the FCC can fight Comcast again over the BitTorrent issue. But the court left it open for the FCC to try other actions to enforce rules on Internet operators. Ultimately, I think the FCC should take a hint from the court and stop trying to regulate the actions of telephone and cable companies at the IP layer. The hint is to regulate them at the level where the FCC has more authority--on the physical level, where telephone companies are regulated as common carriers and cable companies have requirements to the public as well.

The court noted (on pages 30 through 34 of its order) that the FCC missed out on the chance to make certain arguments that the court might have looked on more favorably. Personally and amateurly, I think those arguments would be weak anyway. For instance, the FCC has the right to regulate activities that affect rates. VoIP can affect phone rates and video downloads over the Internet can affect cable charges for movies. So the FCC could try to find an excuse to regulate the Internet. But I wouldn't be the one to make that excuse.

The really significant message to the FCC comes on pages 30 and 32. The court claims that any previous court rulings that give power to the FCC to regulate the Internet (notably the famous Brand X decision) are based on its historical right to regulate common carriers (e.g., telephone companies) and broadcasters. Practically speaking, this gives the FCC a mandate to keep regulating the things that matter--with an eye to creating a space for a better Internet and high-speed digital networking (broadband).

Finding the right layer

Comcast v. FCC combines all the elements of a regulatory thriller. First, the stakes are high: we're talking about who controls the information that comes into our homes. Second, Comcast wasn't being subtle in handling BitTorrent; its manipulations were done with a conscious bias, carried out apparently arbitrarily (rather than being based on corporate policy, it seems that a network administrator made and implemented a personal decision), and were kept secret until customers uncovered the behavior. If you had asked for a case where an Internet provider said, "We can do anything the hell we want regardless of any political, social, technical, moral, or financial consequences," you'd choose something like Comcast's impedance of BitTorrent.

And the court did not endorse that point of view. Contrary to many headlines, the court affirmed that the FCC has the right to regulate the Internet. Furthermore, the court acknowledged that Congress gave the FCC the right to promote networking. But the FCC must also observe limits.

The court went (cursorily in some cases) over the FCC's options for regulating Comcast's behavior, and determined either that there was no precedent for it or (I'm glossing over lots of technicalities here) that the FCC had not properly entered those options into the case.

The FCC should still take steps to promote the spread of high-speed networking, and to ensure that it is affordable by growing numbers of people. But it must do so by regulating the lines, not what travels over those lines.

As advocates for greater competition have been pointing out for several years, the FCC fell down on that public obligation. Many trace the lapse to the chairmanship of Bush appointee Michael Powell. And it's true that he chose to try to get the big telephone and cable companies to compete with each other (a duopoly situation) instead of opening more of a space for small Internet providers. I cover this choice in a 2004 article. But it's not fair to say Powell had no interest in competition, nor is it historically accurate to say this was a major change in direction for the FCC.

From the beginning, when the 1996 telecom act told the FCC to promote competition, implementation was flawed. The FCC chose 14 points in the telephone network where companies had to allow interconnection (so competitors could come on the network). But it missed at least one crucial point. The independent Internet providers were already losing the battle before Powell took over the reins at the FCC.

And the notion of letting two or three big companies duke it out (mistrusting start-ups to make a difference) is embedded in the 1996 act itself.

Is it too late to make a change? We must hope not. Today's court ruling should be a wake-up call; it's time to get back to regulating things that the FCC actually can influence.

Comcast's traffic shaping did not change the networking industry. Nor did it affect the availability of high-speed networks. It was a clumsy reaction by a beleaguered company to a phenomenon it didn't really understand. Historically, it will prove an oddity, and so will the spat that network advocates started, catching the FCC in its snares.

The difficulty software layers add

The term "Internet" is used far too loosely. If you apply it to all seven layers of the ISO networking model, it covers common carrier lines regulated by the FCC (as well as cable lines, which are subject to less regulation--but still some). But the FCC has historically called the Internet a "service" that is separate from those lines.

Software blurs and perhaps even erases such neat distinctions. Comcast does not have to rewire its network or shut down switches to control it. All they have to do is configure a firewall. That's why stunts like holding back BitTorrent traffic become networking issues and draw interest from the FCC. But it also cautions against trying to regulate what Comcast does, because it's hard to know when to stop. That's what opponents of network neutrality say, and you can hear it in the court ruling.

The fuzzy boundaries between software regulation and real-world activities bedevils other areas of policy as well. Because sophisticated real-world processing moves from mechanical devices into software, it encourages inventors to patent software innovations, a dilemma I explore in another article. And in the 1990s, courts argued over whether encryption was a process or a form of expression--and decided it was a form of expression.

Should the FCC wait for Congress to tell it what to do? I don't think so. The DC Circuit court blocked one path, but it didn't tell the FCC to turn back. It has a job to do, and it just has to find the right tool for the job.

Benjamin Smedberg [ 6 April 2010 02:05 PM]

Or of course, congress could just explicitly clarify whether they wish to give the FCC the power to regulate internet providers in this way.

Dave [ 6 April 2010 09:22 PM]

It's the seven layers of the OSI Networking Model.

Obviously you have never studied for a networking certification or you would have known better.

http://www.javvin.com/osimodel.html

Andy Oram [ 7 April 2010 04:57 AM]

Dave: Thanks for the correction. The ISO is the organization that developed the OSI model, so it's an easy slip to make when someone is writing fast to get out an idea when news is hot. You're right: I haven't gotten a networking certification and I don't plan to apply for networking jobs.

Benjamin: You also have a reasonable point, and I know that some congresspeople like Senator Al Franken have talked about submitting legislation--but how long would that take and how many barriers would it have to pass, and how successful could laws be at distinguishing the abuses that require regulation from normal activities that networks need?