Net neutrality, part II

Many readers have been asking me to clarify my stance on net neutrality.  Here are a few qiuck but key points:

1. I favor net neutrality in the current environment.  Without neutrality, Comcast and Verizon would use differential pricing schemes to extract more revenue and thus diminish some forms of Net output, including Google, Amazon, ebay, and possibly blogs.

2. If the cable and telecom companies had no legally-backed monopoly powers, I would not favor legally enforced net neutrality.  "Let the market decide" would be a good answer.

3. Those powers are eroding with time but still the market for high-speed connections is far from contestable.  Municipal wireless would matter a great deal but that is not a pure market solution either.

4. Ex ante, it is hard to predict what will "stick" on the Net.  I see positive and uninternalized social value in the level of experimentation which we currently enjoy.  Profit-maximizing pricing from Verizon and Comcast would choke this off to some degree.

5. Bandwidth might become so scarce that differential pricing is needed to give companies the incentive to create more of it.  Then net neutrality could be a bad idea, even in spite of #1-4.  But we are not there now and maybe we will never be.  Municipal wireless, or some related idea, probably will arrive first.  In the meantime stop watching those silly videos.

6. A related question is this: we all know that road pricing can make economic sense.  But should we favor differential and profit-maximizing pricing on non-contestable roads?  At low levels of congestion, probably not.  It is better to let people get to work.

Here is my previous post on net neutrality.  Here is one good summary of the issues.



How does your solution deal with traffic from low-income geographies that is unable to pay US going rates?

To me it boils down to this:

The best employees at Yahoo, eBay, Google, and other major content providers are engineers.

The best employees at Comcast, Cox, SBC and other access providers are lobbyists.

For both of these types of companies, everyone else they hire is just good enough to support and monetize the work of their best employees.

Net neutrality is actually a lot more complicated issue than either side generally acknowldges.

The current class of service most people use most of the time is moving data where you want it fast on the human scale, but on the network time scales you are very forgiving of delivery time. If a part of the web page you are looking at gets delayed by 300ms, it isn't a big deal. If you loose 15 or so packets somewhere in the network, fear not, tcp will recover.

There are lots of services that could benefit greatly from other, more stringent classes of service. VoIP is a classic that does better in lower jitter, lower packet loss environments. Many such applications are actually pretty low bandwidth all things considered (a voice call is about 64kbps), but they are quite sensitive to the class of service they get. Providing such classes of service costs money, and thus will not happen without market incentives.

Then there are new *kinds* of network services, like multicast. Multicast allows a sender to send out one stream of content which is then replicated by the network to a large number of receivers. It is a much more efficient way to utilize the network for some kinds of network traffic, There are lots of innovative uses for multicast, but the end to end support for it just isn't there on the internet. Again, providing multicast costs money, and no one has figured out yet how to make it pay.

Done wrong, net neutrality rules out these new classes and kinds of network services. While a great deal of internet innovation comes from the edge, and we don't want to choke that off, we also don't want to choke off innovation in the network itself that enables interesting new features on the edge.

So we need to talk a bit more about what we want as an outcome from 'net neutrality'. Most simply, consumers and producers don't want to be deprived of what they are getting now. Also, we don't want strong barriers to entry put up to new entrants offering new kinds of services. Additionally we don't want consumer edge network providers leveraging their near monopoly to edge out other providers of services they choose to tie to their broadband service. Nobody wants their local cable company to be able to screw up their vonage. I also don't want my cable modem provider preventing me from getting video content from other providers (in no small part because my cable providers provides very little video content I actually *want*).

I'm not sure what mechanism is sufficient to acheive these goals. In a healthy market place, I'd say trust the market. We don't have a healthy market place in broadband in most places. It's pretty much tweedle dee or tweedle dum (cable or dsl), and I'm frequently as a loss to tell you which is more evil, stupid, and incompentent :)

Ben M,

You are oversimplifying the technical problem. Consider a voice call. First, comcast would have to identify what of my traffic need the low latency, high reliability, low jitter transport. Then comcast would have to identify what of my in bound traffic (the inbound voice stream from the other party) requires that class of service. When comcast provides you with this service they use a particular sub set of the DOCSIS protocol to provision that class of service, but that's not designed to allow arbitrary third parties to request prioritized streams (you, the customer being a third party). It's actually pretty complicated to arrange a system where you can let the consumer dynamically request prioritized service for their realtime traffic.

Then consider I am pulling down some streaming video I'd like prioritized. I click on the link... how exactly do we tell the network to prioritize that traffic? The protocols to do so aren't even there, much less the rest of the infrastructure.

From my point of view, I think end user driven classes of service are the way to go, but we are a long way from the necessary infrastructure for that. Given the near monopoly nature of the consumer edge ISPs, their are very unlikely to go that route unless forced (either by government or competition, and I don't see much competition for them). If you want to try to do it by regulation, you need a decent definition of what you mean by network neutrality, and I haven't seen one yet.


Normally, I'd agree with you about anti-trust law being the right approach. Unfortunately, Microsoft has pretty conclusively proven that anti-trust law is a dead letter these days. The market is moving so fast that you can abuse your monopoly power, crush your competitors, scorch the landscape completely, and move on, all before Justice even notices there may be a problem, much less files and anti-trust suite. Further, you can simply agree to a consent decree, ignore it, and be pretty certain the consequences will be of little importance to you.

I'm sorry to say I've lost faith in anti-trust law at this point. That's why folks are seeking a proactive net neutrality solution.

Tyler -- You write here that: "If the cable and telecom companies had no legally-backed monopoly powers, I would not favor legally enforced net neutrality." But, today, neither cable nor telco's have legally-backed monopolies. Monopoly franchising for telephone companies was banned by the 96 telecommunications act. The 92 cable act similarly banned exclusive franchising for cable service. Of course, local authorities have made it difficult for potential competitors to actually get franchised -- but their ability to do this would be eliminated by the telecom bills now moving through Congress. So I'm puzzled by your reference to legal monopolies. Am I missing something?

"I believe Tyler is referring to the natural monopoly of telephony and cable wiring. It makes no sense to have a truly competitive market for internet pipe, for the same reason that truly competitive markets for electricity aren't efficient."

Robert -- two problems with this. First, Tyler said "legally-backed" monopoly, which is far different than the type of economic monopoly you describe here. Moreover, whatever the situation in the past, there certainly is no natural monopoly in telephony or cable today. Competition not only is possible, but it is here. Today. Right now. Cable is making great inroads into voice calls (using their own facilities), as are wireless firms. In cable, satellite provides cable today, and telco's soon will. And in broadband itself, both telcos and cable (in addition to a number of emerging competitors, offer service to consumers.

I know there are lots of theoretical arguments out there as to why competitive facilities in these markets should never exist. Yet, someone forgot to tell the marketplace, because those facilities -- and competition -- are nevertheless there.

"It's difficult to gather hard evidence for these cases, as Comcast could (and does) claim that line maintenance and other factors caused voIP outages."


Why is it difficult to gather hard evidence? I would think that a systematic degradation of service aimed at a specific type of traffic should be easy to measure. Surely Vonage has the ability and incentive to investigate a serious allegation like this if it's credible?

What does "I favor net neutrality in the current environment" mean? Does it mean you would support legislation compelling all companies that sell Internet access to the general public to refrain from deliberately providing differential service based on destination? Or what?

"If the cable and telecom companies had no legally-backed monopoly powers, I would not favor legally enforced net neutrality." So you plan to force net neutrality only on the cable and telecom companies as an exchange for their legally-backed monopoly powers? Or would you also enforce it on those companies already struggling to compete with legally-backed monopolies?

What is making me, and it seems quite a few others, inclined more toward the neutrality camp is the question of what the lack of common carrier status will mean for the ability of ISPs - or any backbone router owners for that matter - to decide for the rest of us what content is worthy of transmission at all, extra fees or not. It seems that these companies may shoulder the mantle of the nanny state themselves, all the while decrying any potential recourse from "big bad government" or organized consumers to this state of affairs. It's a pain, but the cost of broadband service is still just expensive enough that I can keep my dial-in line a bit longer. Or perhaps I should just look at it this way: AT&T's dial lines are common carriers, and their broadband is not. If they pull shenanigans with their broadband, I can just resort to using my dial-in line to access providers they don't like. The only problem is they charge a termination fee if you shut it down within a year which basically makes it cheaper to stay for the whole year if you've already been there for 4 months. Comcast meanwhile has no termination fee but has an installation fee if, like myself, you don't already have cable. I hear people talking about setting up workarounds on 802.11, but I fear we may see a split internet as this goes on, a ragtag radio conspiracy trying to build around the central network a-la The Matrix. And of course, the land-line providers would be plenty happy to fund (if they aren't already) groups who speak on the health hazards of E/M radiation - though I'd dearly love to know if there's ever been a study on CB or amateur users which shows the degree of risk they've suffered.
It is a complicated question, this future we are building.

Hello, nice site look this:
End ^) See you

Hello, nice site look this:
End ^) See you

Comments for this post are closed