I’ve been meaning to do another post on net neutrality based on a fair amount of recent activity on various blogs but haven’t had time. To cut down on the number of open tabs in my browser, I’m just going to dump the links here and let you read them yourselves:

• Vint Cerf for Google on reasonable approaches to managing networks
• A “history lesson” from Verizon
• Joost CEO Mike Volpi on net neutrality and Adam Smith

I’m not convinced any of these companies has 100% the right approach but I’m glad we’re finally discussing it in a reasonable manner rather than simply posturing or suing one another about this…

So the FCC finally dealt with the Comcast net neutrality issue and issued a 3-2 decision castigating Comcast for doing whatever it did but not fining it:

Ruling on a complaint by Free Press and Public Knowledge as well as a petition for declaratory ruling, the Commission concluded that Comcast has unduly interfered with Internet users’ right to access the lawful Internet content and to use the applications of their choice. Specifically, the Commission found that Comcast had deployed equipment throughout its network to monitor the content of its customers’ Internet connections and selectively block specific types of connections known as peer-to-peer connections.

The problem is, these “rights” don’t exist - at least as a fact of law. Neither the FCC nor Congress has ever passed a ruling or legislation that says Comcast can’t do what it did. There have been various policy statements - especially the famous one about the four net freedoms - but those don’t carry the weight of law. So the FCC is essentially finding Comcast in contravention of a law and passing the law at the same time, which is exactly Comcast’s complaint, and one that’s echoed by the two Republican Commissioners - here’s Commissioner McDowell:

In short, we have no rules to enforce. This matter would have had a better chance at appeal if we had put the horse before the cart.

That’s an important secondary point - this will likely end up being simply a symbolic gesture because the appeals courts are bound to side with Comcast and the two Republicans - there was no law there to enforce and so Comcast had no way of knowing what it did was wrong. Which means as well as falling short in this particular case, it will also fail to set the precedent this is already being hailed as.

Now, I’m no shill for Comcast. As my previous posts on this topic show, however, I do have a certain amount of sympathy for its position. I do think it’s far more reasonable for a cable company to adopt traffic management policies to deal with occasional network congestion than to invest in a massive upgrade of its network to ensure that congestion never occurs. However, according to the evidence the FCC majority accepted to be true, Comcast throttled all P2P traffic all the time, regardless of congestion, which is unreasonable even if you think they have a right to throttle traffic when there is congestion. It’s not clear the FCC did any of its own research on this topic - it would have been hard to, since Comcast had discontinued the practice some time ago - but it seems to have taken various consumer rights groups’ word for it.

I still think the most reasonable approach to all this is to give ISPs - whether cable companies, telcos or wireless operators - the right to set their own reasonable use terms and make these clear to customers, including whatever network management policies the provider intends to apply in general terms, and then let customers decide whether they’re willing to put up with those or not. As I’ve mentioned before, this already applies in the case of running servers on consumer DSL lines, for example, and no-one seems to have objected to that.

However, it appears Kevin Martin has decided to step away from the set of principles he held when he first took this position and go with the popular flow as a bid to leave some kind - any kind - of legacy before he gets ditched as one of the first acts of an Obama presidency. The problem is that the decision is both wrong and unenforceable and so it will end up being remembered as a hollow gesture that did very little real good.

This is intriguing. One of the “inventors of the Internet” (I’ve heard this title applied to quite a range of people, Al Gore included, of course), Lawrence Roberts, is involved in a company which is providing network boxes that throttle P2P traffic in order to allow other traffic to flow freely. So far:

You’ll find Anagran bandwidth fairness boxes (also called FR-1000s) in university settings now, where the P2P file transfer problem is most acute. Anagran doesn’t currently have any commercial ISP customers, but I’ll bet that they’re all looking at them.

Absolutely, and no doubt Comcast is among them. I still maintain that the best approach to dealing with P2P traffic is having an explicit public statement about your policy towards it that all your customers can read and be aware of, and then throttle/shape the traffic in such a way that other forms of traffic are unaffected during periods of network congestion. At any rate, this is another wrinkle in the always interesting net neutrality debate.

There’s an article in Monday’s Wall Street Journal by Andy Kessler, who is a former hedge fund manager and writes books. And it’s a good example of how not to argue against net neutrality.

It’s such a muddled argument that it seems for at least half the time as if he’s arguing for the other side. He gets his facts wrong (suggesting Comcast blocks P2P downloads rather than uploads), makes invidious comparisons (because Comcast’s packet reset technique is the same as used by China to censor content the two are apparently in the same boat) and generally seems to be uninformed about the debate or the current state of the market. He mentions AT&T as a competitor to cable, but only as a provider of DSL services (he either doesn’t know about or conveniently forgets to mention its fibre rollout, which is delivering substantially faster speeds), and he doesn’t seem to know about Verizon, which is rolling out fibre too.

His solution is to drive more competition, which would apparently stimulate a rollout of fibre (ahem - see previous paragraph). Qwest is also rolling out fibre to the node in 23 of the markets it serves in 2008. He does seem to acknlowedge this fact later on but only in the context of this paragraph:

Municipal or privately run wireless data services using Wi-Fi or WiMax should be sprouting like weeds. But they aren’t being built because of lack of access to street lights, of all things, to set up access points. Verizon is busy rolling out a fiber optic service, FIOS, that will provide much higher speeds and real competition to Comcast. But it is slow going, as state by state video franchise rules still favor cable over any newcomers.

He seems to be unaware that several attempts have been made to run municipal WiFi and on the whole they’ve been flops, not because of access to street lights (which, in municipal deployments, aren’t a problem) but because the technology and business model are lousy. WiMAX is being deployed by Towerstream and Clearwire among others, with Sprint set to follow, but it too is unproven as a technology which would compete effectively with fibre rollout. And his details on FiOS are also a little out of date - it’s been some time since Verizon complained about franchising, since it’s actually making very good progress on that front with help from national and state-wide efforts to ease franchising processes.

My favourite line, though, has to be this one:

We have faux competition, cable monopolies versus phone monopolies

Does anyone want to volunteer to help Mr Kessler with a definition of the word monopoly?

Another great set of paragraphs:

A stroke of a pen can cure these ills, incumbents be damned. They will adjust. I personally would climb telephone poles on my street to run fiber if I could get 100 megabit Internet service. Any takers? Talk about an economic stimulus; this is the type of infrastructure we need. The stock market will fund it all as well as resolve overbuild problems.

Don’t think of Internet access as a static business — someone put in phone lines 50 years ago or cable lines 20 years ago, and we are stuck with their limitations. Technology changes the game every few years. Even fiber lines put in today will be obsolete within 10 years and need upgrading. Same for wireless systems.

The trick to an open and innovative Internet is not sneaky technical fixes nor more rules and regulations and bureaucracies to enforce them. The Internet will only expand based on competitive principles, not socialist diktat. The more we can do to clear a path, the greater our national wealth will be. Comcast did us a favor by bringing this net neutrality debate out in the open. I hope the FCC doesn’t fall for this lousy idea.

I love the way that he’s proposing massive new regulations on incumbents in the form of rights of way and pole access, but also decries “socialist diktat”. What’s the difference, Mr Kessler? If the government should be getting out of the way, then let them get out of the way. Don’t go backwards to where we were two to 12 years ago, when the kind of access he’s proposing was part of the failing regulatory regime. Competition is thriving. Yes, in most areas it’s currently a duopoly between cable and telco, but wide area wireless broadband through the cellphone companies and WiMAX-based providers are potential competitors too, and all because government finally got out of the way and stopped trying to create competition through regulation, allowing the market to work.

Nice to see there are some influential figures within the world of the Internet who have a different view about net neutrality. George Gilder and Bret Swanson had an article in the Wall Street Journal this week about their views on net neutrality and they closely mirror my own.

Some key quotes:

We need a dramatic expansion in raw capacity, or bandwidth, and also fine-grained traffic management capabilities to ensure robust service for increasingly demanding consumers. But none of this can happen if we regulate complex network traffic engineering and experimental business plans.

…
Capacious, big-bandwidth networks will transcend many of today’s specific complaints. As raw capacity expands, more and more applications and users can peacefully coexist. But inevitably, sophisticated network users with innovative applications will find creative ways to push the boundaries of capacity on certain network links, and some bits will be shuffled and queued.

The network is now a global computer made up of hardware, software and human minds. But this new, fast-changing and highly organic computer is no more easily regulated than were the circuits, storage, memory and protocols of a mainframe or PC. Leaving it to Washington agencies and committees to engineer the exaflood would be an act of unimaginable folly.

This nicely mixes the practical reasons for avoiding total net neutrality (the present and ongoing need for traffic management to deal with bandwidth-intensive applications) and the philosophical idea that government and legislation is the wrong solution for almost anything having to do with the way the Internet runs. It’s simply too fast moving, as the 1996 Telecom Act has abundantly proven.

Comcast on Tuesday filed with the FCC its response to complaints by Vuze and Free Press about its traffic and network management practices. It’s an 80-page document, with the first 60 pages or so outlining Comcast’s arguments and the last 20 pages attaching its latest acceptable use policy and FAQs regarding excessive use and network management.

It’s worth looking at in detail because Comcast makes some good arguments and some which are a little flimsier.

At root, Comcast’s argument is that:

• it has (so far) restricted its network management activities to the following narrowly-defined set of circumstances: uploads by BitTorrent users who are not simultaneously downloading (and are therefore thought to have left their PCs unattended as peers in the network rather than as active consumers of torrents), when networks are congested
• it has sufficient capacity in its network to avoid congestion when it monitors only this very specific type of traffic, but when these uploads are permitted it faces congestion problems
• network management is therefore the only way to ensure that all other traffic can reach its destination at times of high usage of the network
• the FCC ought to rule that such network management is “reasonable” and in fact the kind of thing that a responsible ISP ought to be doing to provide quality service to customers
• there is no merit in its opponents that it has done something illegal, since the FCC has never actually turned the much-touted net freedoms espoused by Michael Powell into anything with legal authority.

If you assume (which some of its opponents won’t) that Comcast is being honest and open, this seems to me to be entirely reasonable. If one very narrow and specific category of traffic is causing all the problems, and temporarily blocking that form of traffic during peak times solves the congestion problem, it is absolutely the right thing to do. It is the least intrusive way of managing the problem, and affects only traffic which users generally have no real interest in anyway (since the content is flowing away from users, usually in an automated fashion, and therefore is not content they are consuming). Certainly, there is no indication that the content which is being blocked in any way directly competes with any service offered by Comcast, since it is not blocking downloads.

Is there anything to object to at all here, then? Well, yes. Comcast has taken a very long time to release this data about its network management activities, and arguably has obfuscated the truth throughout the process. Now, in this it is no different from other ISPs out there, none of whom provided detailed information about their network management practices either. As Comcast itself says in its filing, requiring such a level of detail to be filed would have two adverse effects: (1) it would require significant manpower to always have the latest policies posted somewhere, (2) it would provide information to both legitimate service providers and those with nefarious motives about how to bypass its controls, defeating the object of network management and only increasing the burden on Comcast.

What are the alternatives to network management? It seems to me there are essentially two of them:

• allow degradation to take place at busy times
• massive investment in additional bandwidth.

The former would cause a worsening experience for all customers and applications at peak periods, which would likely lead to complaints against Comcast and customer desertions. It would technically be “neutral” and “non-discriminatory”, but wouldn’t really be in the best interests of any particular party - consumers or application or service providers. The latter would be hugely expensive in the first place, and up to a certain point any additional bandwidth would likely be scooped up by P2P activity anyway, which means it wouldn’t solve the problem.

All of this leads me to wonder whether the ISPs - Comcast included - wouldn’t be better off simply banning the use of P2P software (or at least P2P uploads) in their acceptable use policies. Here are some excerpts from other AUPs:

You agree that the Service is not to be used to host peer-to-peer applications that you are not actively using [from AT&T's terms of service]

You may NOT use the Service as follows: … (x) to install “auto-responders,” “cancel-bots” or similar automated or manual routines which generate excessive amounts of net traffic… You may not use the Broadband Service to host any type of server whether personal or commercial in nature. [from Verizon's series of tubes” analogy. Speaking of which, Comcast relies on a quote attributed to Congresswoman Mary Bono Mack (widow of Sonny Bono and inheritor of his House seat) in describing BitTorrent:

As Congresswoman Mary Bono Mack recently explained:

The service providers are watching more and more of their network monopolized by P2P bandwidth hogs who command a disproportionate amount of their network resources. . . .

You might be asking yourself, why don’t the broadband service providers invest more into their networks and add more capacity? For the record, broadband service providers are investing in their networks, but simply adding more bandwidth does not solve [the P2P problem. The reason for this is P2P applications are designed to consume as much bandwidth as is available, thus more capacity only results in more consumption.

As others have pointed out, Congresspeople are never the strongest authorities on such topics and it probably could have found better ones. An article quoted later cites this snippet from Bram Cohen, the founder of BitTorrent, and actually does the job just fine:

Cohen agrees. In fact, it's something he predicted when he first thought up BitTorrent. "My whole idea was, 'Let's use up a lot of bandwidth,'" he laughs. "I had a friend who said, 'Well, ISPs won't like that.' And I said, 'Why should I care?'"

At the end of the day, the salient fact is that a small number of BitTorrent streams (Comcast suggests as few as 10 or 15) in a single node can start to cause problems, and a single user can easily be responsible for most or even all of those. Strangely, Comcast offers no evidence of its own to support any of its claims. It must have masses of data on the amount of P2P traffic in its network and so on, from the Sandvine servers which it uses to filter traffic in the first place. Perhaps it assumed that no-one would take its word for it.

One trap Comcast doesn't fall into is using the illegality of much of the file sharing that BitTorrent enables as an excuse. It doesn't mention this once, to its credit, thereby negating the obvious counter-argument that not all BitTorrent uses are illegal. It is, however, guilty of making statements which are stronger than the supporting evidence cited in the filing itself entitles them to be:

In no event does Comcast prevent, restrict, or limit the use of applications and services using P2P protocols

Given that the most logical meaning of this statement is clearly untrue from everything it says earlier in its filing, this is again disingenuous to say the least. The following statements are somewhat more accurate, in that Comcast doesn't prevent consumers from using certain applications altogether.

Comcast’s customers have unfettered access to any lawful content they choose, including content that is delivered via P2P protocols. Comcast’s customers’ P2P downloads are utterly unaffected by its network
management practices, and the limited network management measures applied to certain P2P uploads in certain circumstances are entirely content-agnostic [my emphasis]

All of this makes you wonder at what point Comcast’s management kicks in. Based on the figures it quotes about 15 or even 10 BitTorrent streams in a single node causing problems, it might be restricting the number of unidirectional uploads to less than 10 in any given node at any given time. Of course, some streams are bigger than others, and there are also bidirectional uploads (i.e. people uploading while downloading) which Comcast claims not to be managing, so it’s entirely possible that networks could become congested anyway (Comcast seems to suggest this is not the case).

At the end of the day, I’m inclined to agree with Comcast’s reasoning, but it’s undeniable that its handling of this whole issue has been clumsy from start to finish and this filing in particular could have been a lot tighter. Hopefully the FCC will see it the same way, although it may all become moot if Markey’s latest effort gets the go-ahead.

After failed attempts to get net neutrality legislation passed in 2006, Ed Markey and friends are at it again. This time, the bill has been softened a bit, with the language toned down and the focus on giving the FCC more authority and a mandate to enforce three of Michael Powell’s Four Net Freedoms as drawn from an official policy document of the FCC.

The Bill is actually presented as an amendment to the Communications Act of 1934, as was the 1996 Telecom Act. The preamble lays out the purpose of the Bill, which seems inocuous enough:

The importance of the broadband marketplace to citizens, communities, and commerce warrants a thorough inquiry to obtain input and ideas for a variety of broadband policies that will promote openness, competition, innovation, and affordable, ubiquitous broadband service for all individuals in the United States.

It then moves into rather shakier territory:

It is the policy of the United States… to maintain the freedom to use for lawful purposes broadband telecommunications networks, including the Internet, without unreasonable interference from or discrimination by network operators, as has been the policy and history of the Internet and the basis of user expectations since its inception.

Here’s where things get tricky, because in proposing a new set of formal rules for the Internet Markey is relying on a presumed “policy” that has existed hitherto, but this of course is nonsense. Since the Internet is a private enterprise made up of thousands of individuals and companies, it hasn’t had a policy because it hasn’t needed one - it’s been run according to the competing priorities and goals of all those involved in it.

The other crutch for his position is “user expectations” which, of course, is at least as vague, if not more so. And herein lies the problem - rather than allowing the FCC full freedom to determine what future policy around the Internet should be, what user expectations really are, and how to balance those against the real-world economics that underpin the Internet, Markey has taken the first few steps for them and thereby limited their options. So, far from allowing the FCC to undertake “a thorough inquiry to obtain input and ideas for a variety of broadband policies” Markey has provided a much narrower framework than his preamble suggests.

A little lower down in the policy section the Bill states that it should be US policy:

(3) to preserve and promote the open and interconnected nature of broadband networks that enable consumers to reach, and service providers to offer, lawful content, applications, and services of their choosing, using their selection of devices, as long as such devices do not harm the network; and

(4) to safeguard the open marketplace of ideas on the Internet by adopting and enforcing baseline protections to guard against unreasonable discriminatory favoritism for, or degradation of, content by network operators based upon its source, ownership, or destination on the Internet.

Part (3) is merely a summary of the three net freedoms which will be enumerated later on, but note the reference to legality here. Part (4) then goes on to state the simple premise of neutrality, which forbids “unreasonable” treatment based on source, ownership or destination. For starters, we have a conflict between the ownership provision here and the legality provision in part (3) - if a provider was able to conclude that content being shared was not owned by the person sharing it and therefore was being transmitted illegally, would that come under part 3 (where it would be allowed) or part 4 (where it wouldn’t)?

And how do we define “reasonable”? Comcast (the only provider to have openly stated it plans to be non-neutral in its treatment of traffic) would argue that its traffic management and shaping are reasonable because they are necessary to allow its services to all customers to run smoothly without massive additional investment. On what basis will the FCC second guess this judgement if the only criterion is that such action should be “reasonable”?

The specific mandate for the FCC to look into the current state of the market and determine appropriate policies comes a little later:

Within 90 days after the date of the enactment of this Act, the Federal Communications Commission (in this Act referred to as the ‘‘Commission’’) shall commence a proceeding on broadband services and consumer rights.

Specifically, this proceeding needs to determine the following:

• Whether any provider is currently in violation of the three net freedoms (relating to applications, content and devices)
• Whether “broadband network providers add charges for quality of service, or other similar additional fees or surcharges, to certain Internet applications and service providers” in contravention of the net neutrality principles set out earlier. This obviously goes right to the heart of one of the two scenarios which have been laid out by providers under which they would want to be less than neutral - to deliver content at premium speeds and quality for a premium price (the FedEx vs. USPS model), and is therefore worrying. In addition, it’s not clear how this would affect players like Akamai, which specialise in providing such services.
• Whether “broadband network providers offer to consumers parental control protection tools, services to combat unsolicited commercial electronic mail, and other similar consumer services, the manner in which such services are offered, and the extent to which such services are consistent with such policies of the United States”. Most people seem to have read this as a sop to Republican lawmakers but I wonder if it isn’t actually the opposite - since it’s hard to imagine a way in which the lack of parental controls and spam filters could contravene net neutrality, but it’s possible to see how an extreme interpretation of net neutrality principles could lead you to believe the presence of such controls would violate them. At the very least the intent of this paragraph needs to be clarified.
• Network prioritisation practices by providers and their consistency with net neutrality principles
• The relationship between net neutrality principles as applied so far with competition in Internet services
• Whether providing a sufficiently fat pipe could excuse a provider from these rules. This is a particularly interesting one - the whole point is that we’re getting fatter and fatter pipes in the access network, but the backhaul portion is where the bottleneck occurs, and that’s the very reason for wanting to offer QoS enabled services in the first place. Carriers want to avoid building out massive additional bandwidth in that part of the network by more effectively prioritising traffic - few will qualify for this condition as a result, unless the FCC is stupid enough to look at access speeds alone. But it also goes to the heart of the matter - whether prioritising some traffic necessarily degrades all other traffic, including competing traffic. If it does, the argument for neutrality at least has some merit. If it doesn’t, then that argument essentially goes away.

The Bill further requires eight additional broadband summits to be held around the country within one year of the passage of the Bill into law. These summits will exist

to assess competition, consumer protection, and consumer choice issues related to broadband Internet access services.

Note no mention of current limitations of the Internet, the economics associated with the current business model and any future business models either in keeping with or in contravention of net neutrality. And the “technology sector” is the last group mentioned in the list of individuals and entities to be invited to such summits, and presumably includes providers, though that’s not explicit. The focus is very much on consumers and their “rights” and views rather than on a balanced evaluation of the competing interests of consumers and providers.

The output of all this is a report back to Congress with recommendations, which is the kicker here. Representative Chip Pickering (R) has signed on for this Bill but had opposed previous attempts, on the basis that this one was rather softer than previous attempts in its aims and prescriptiveness. However, the idea that the ball is going to be back in Congress’s court rather than remaining with the FCC at the end of the process suggests this may just be a smokescreen to get a net neutrality Act into the statute books and then follow up in a year’s time with a much harsher version based on all the work done by the FCC. The Bill has largely been described in blogs and articles as giving the FCC more teeth to deal with abuses of the net freedoms but there’s actually nothing in here that does that. It merely asks the FCC to go and do the legwork and then hand the results over to Congress for a solution. At a time when Democrats (who largely strongly favour net neutrality) are likely to increase their majority in Congress and also appoint a Democrat-led FCC, this is particularly worrying.

My own position on net neutrality has always been this: that those willing to pay a premium for premium delivery (whether they be content owners or end users) should be able to do so, and that the best way to architect such a delivery system is to do it in parallel with the existing Internet infrastructure such that other traffic is not degraded by it. To the extent that there is a bottleneck in traffic, all remaining traffic would be degraded equally by virtue of its volume, not because of any discriminatory measures. I also think it would be reasonable within that broad flow of content to prioritise real-time traffic (primarily voice and video) and allow non-real-time traffic such as website data and emails to be processed slightly more slowly. I don’t think the Bill as currently written would allow that model to exist, and I therefore believe it’s deeply flawed.