by Michael Cain
This past week the House Subcommittee on Communications and Technology met to discuss a draft bill on network neutrality. A Senate panel will meet next week to discuss the same bill. Congress is very early in its process; the bill is a “discussion draft” and nothing has been formally introduced yet. What I’m going to argue in this post is that the bill is actually a pretty reasonable compromise, once all things are considered.
Some technical background (I’ll be brief)… The internet works by sending packets of data between different computers (some dedicated to the job of passing packets). Think of a packet as an envelope — it has a destination address, the sender’s return address, and contains a chunk of data. The internet doesn’t care about the data, it just moves the packet based on the addresses. Technically, the promise that each computer on the internet makes to the sender is “I’ll try to make an attempt to send your packet to another computer that I think is closer to the final destination than I am.” (That enormously reliable service can be built on such a weak promise should amaze everyone.) Traffic shaping is when the internet service provider (ISP) modifies that promise based on the source and/or destination addresses, perhaps because someone paid more for their service. At one time, for example, some ISPs simply discarded packets whose addresses indicated they were being used for peer-to-peer networking, where the content was overwhelmingly involved in copyright violations. Network neutrality (or net neutrality) is a public policy that says ISPs shouldn’t engage in traffic shaping.
Some recent political background… Major ISPs (we’re looking at you, Comcast and Verizon) engaged in traffic shaping. In particular, they limited how quickly packets from Netflix would get delivered, which had an impact on Netflix streaming quality. Then they told Netflix, we’ll take the limit off if you pay us. The FCC issued rules saying such behavior was not allowed. Comcast and Verizon challenged the rules in court, on the principal that the FCC had classified internet access as an information service 20-some years ago, and information services are exempt from that sort of regulation. The SCOTUS (rightly) agreed with Comcast and Verizon. President Obama said that the FCC should reclassify internet access as a communication service that can be regulated that way (he can’t simply order it done, the FCC is an independent agency, and there’s a process for changing the rules). The FCC had previously agreed to look at other options and will issue its decision in February. Most people anticipate that the FCC will reclassify internet access to be a communications service and then re-issue the rules that forbid traffic shaping.
What the Republican discussion draft bill would do… (1) It makes it illegal for ISPs to implement traffic shaping for reasons other than reasonable network management. Those network management practices must be revealed in advance and be non-discriminatory. (2) It requires the FCC to create a formal procedure for handling traffic shaping complaints. That’s necessary, since the FCC is getting a new power. (3) It allows the ISP to create specialized services, but states that specialized services may not be used to evade the purposes of the bill. (4) It makes internet access an information service permanently, thereby blocking FCC actions on reclassification (or reversing them, if it had already happened).
Most of the published commentary — heck, all of the commentary I’ve read — is critical of the draft. On the one side are those who back the traditional Republican position on the subject, and assert that competitive markets will deal with the problem — that if consumers don’t like traffic shaping, they’ll take their business to ISPs that don’t do traffic shaping — so there should be no regulation at all. On the other side are those who claim the bill doesn’t go far enough, and that the “special services” provision will turn out to be a loophole that allows ISPs to do traffic shaping anyway. On the third side (and I’ll avoid the appropriate but obscure “gripping hand” reference) is me: I claim that the bill is actually a reasonable compromise, and that both the other sides of the commentariate are missing a really major point.
Traffic shaping is remarkably unpopular. The FCC had to extend the public comment portion of the process leading to next month’s decision because the volume of comments broke their comment system. The public comments are overwhelmingly in favor of banning traffic shaping. Tech companies like Amazon, Google, and Microsoft are overwhelmingly in favor of banning traffic shaping. The “special services” language is intended to grandfather in services like Comcast’s voice telephony, which gives priority treatment to voice packets on the local distribution system only. There are valid technical reasons why those exceptions need to be made. The Republicans had no real political option except to get on the “ban traffic shaping” bandwagon, and the bill effectively bans traffic shaping.
However, what the other side misses is that reclassification as a communications service opens up all kinds of regulatory headaches besides traffic shaping. They don’t get that because most of them didn’t live through the break-up of the Bell System and the experience of dealing with the Telecommunications Act of 1996 in actual practice. If internet access were reclassified, then eventually the existing cable, telephone, cellular and fiber companies are going to be on the hook for a lot of very expensive network changes. Based on their historical relationships with the big phone and cable companies, the Republicans had no real political option except to keep internet access on the information-service side of the fence.
But I claim that’s a good thing, not just for Comcast and Verizon, but also for Google with its new fiber networks and for municipal networks. If internet access were a communications service, and all of the existing rules for communications services were eventually applied, things get ugly. There are real reasons why the FCC didn’t do the reclassification the first time around. I would pay to watch how Google reacts when a dozen other ISP companies show up with requests to colocate equipment in Google’s head end and be allowed to purchase raw bandwidth on Google’s fiber at the price Google charges itself. What’s that you say, Google? You don’t charge yourself for the bandwidth? Sorry, the rules for communications services require that you do your accounting that way, and in excruciating detail. At that point, if Google had any sense, they’d get out of the business entirely. And the potential ISPs will show up. I was peripherally involved in an experiment at a modest-sized cable company in the late 1990s that explored colocation and bandwidth sharing with other ISPs on the cable network. The kind of resource sharing they demanded would have effectively rendered the system unusable for any ISP, and the experiment was canceled.
All-in-all, I’m satisfied with the bill as a compromise. It does away with traffic shaping; and it doesn’t open up the enormous can of worms that FCC reclassification would. In the long run, we ought to change how internet access is regulated. I said that 20 years ago when the current rules were adopted, and still say it. But that’s a really big job, and will take years while the FCC and the courts redesign various distribution networks. In the short term, fix the traffic shaping problem. That’s good enough for me, for now.