« Welcome Home! | Main | Political Purges and Civil Service »

April 23, 2006

Network Neutrality

There have been a couple of posts by Dale Franks at QandO recently, here and here, dealing with network neutrality (the proposition, or design principle really, that the INTERconnections between NETworks that make up the Internet should not care about the content of the data passing across them — a packet is a packet is a packet). The alternative is "intelligent networks", which look at the source, destination, type or contents of packages and decide to charge differently, or throttle differently, based on those characteristics.

For example, let's say that you have a cable modem, as I do. Your cable company might decide that Google is "using up too much bandwidth" because so many people use it. The cable company would then go to Google and say that traffic to and from Google would be throttled (artificially limited as to the amount that could be sent across the cable company's network) unless Google paid a fee to make up for its "excessive usage" of the cable company's bandwidth. Google would then have the option of striking deals with the cable and telephone monopolies that provide most of the "last mile" broadband to internet users; setting up alternate ways of getting to its users; or being throttled, with the consequent loss of reputation (they will appear to perform badly), traffic and revenue.

Presumably, there would not be such a problem with Google's (or Microsoft's, or Yahoo's, or any other large site's) own provider. For one thing, Google pays for vast amounts of bandwidth directly carried by long-haul carriers over dedicated lines. For another thing, if any of Google's (presumably multiple) providers threatened such a thing, Google could just cancel their contract with the provider and use its other alternatives, some or all of which would be glad to take the additional cash in vast amounts.

What really would be happening is that the local providers would be ignoring the fact that its customers are in fact paying for their bandwidth, and would be attempting to get other, deeper pockets to pay for the provider's bandwidth a second time. From a bottom line perspective, it makes a lot of sense in the very, very short run for local providers to use bandwidth shaping and similar technologies to wring more money out of the network, just as it pays for them to set up added services (like email or web content provision or hosting) that are so bad none of their customers will actually use them, so that their costs of setting up are small but they get extra revenue from charging for the unused services. But in the case of bandwidth shaping, this really is only a very, very short term benefit.

Let's say that my local cable company were to put such limitations in place. I would notice, and would immediately cancel my contract in favor of another local vendor that did not have such limitations. If I could not find such a vendor, I'd lease a T1 line (a dedicated, high-speed line). Even for people who cannot lease a T1, there are usually multiple providers in urban areas, where most broadband consumers live. And there's no percentage in inconveniencing your rural users, because let's face it, there aren't enough of them for bandwidth shaping to be a threat to deep pockets companies. The collective loss of business, possibly combined with efforts to end local monopolies for cable and phone service, would quickly put an end to the business utility of bandwidth shaping. In any but the short term, unless there is a forced monopoly, there is no percentage in inconveniencing your customers. And while there may be a cable monopoly, and a phone monopoly, and both may collude in inconveniencing their customers, there are still other ways of getting high-speed connections which would suddenly become much more attractive. In other words, there is no monopoly on the provision of high-speed IP connectivity even if particular methods of providing high-speed IP connectivity are monopolies.

So while I support Dale's general idea of deregulating telecom services as much as possible, I don't really see network neutrality violations as being much of a big deal: the Internet routes around damage. While it may require building more connections than currently exist (a good thing in any case) in order to do so, the advantage from bandwidth shaping would be so short term as to create no real problems for internet users.

UPDATE: See also here and here. I thought about whether to put my response to Scott Chaffin in the comments or here, and decided to leave it in the comments. But if you agree with Scott that the problem is "free riders", you might read my response in the comments for a concrete example of why this is not so.

Posted by jeff at April 23, 2006 5:36 PM

Trackback Pings

TrackBack URL for this entry:


There are some distinct problems here that nobody has acknowledged yet.

- Upgrading the infrastructure is expensive. As Google and Yahoo et al offer more and more Ajax-type apps, they eat up more bandwidth on the existing infrastructure. Who pays to upgrade that infrastructure to keep the customer happy? Customer of who, exactly?

- A packet is a packet until YouTube and Skype and 50 other social web apps running on your PC interfere with me trying to get my corp email over my VPN. Especially when your Skype session becomes a Skype super-node because Skype are bad programmers.

- Why would any provider build more connections if there is no way to recoup the cost of building them?

I'm a huge web hippy, and I understand what all the net neutrality guys are saying, but it smacks of free riders who are demanding that somebody else keep supporting them. Bandwidth is finite and people spend money to provide it, and not inconsequentially, upgrade it.

Posted by: Scott Chaffin at April 23, 2006 8:04 PM

I pay a certain amount per month to my provider to give me a certain amount of bandwidth. Since I am on a cable modem, I expect the head-end to be shared, and thus the available bandwidth to be shared, and thus the advertised maximum bandwidth to be a best-case estimate.

Google pays a certain amount per month (vast, vast amounts per month, in fact) to its various providers to provide it a certain set amount of bandwidth, and since Google is undoubtedly using dedicated circuits, they can actually expect to get all of that bandwidth.

How does it get from Google's provider to my provider? Well, my provider pays a long-haul carrier like AT&T, and so does Google. At one of the MAE's, or some other colocation point, these connections come together.

In other words, I am paying for my bandwidth to the carrier, and a fraction of their bandwidth over their various long-haul carriers. Google is paying for their bandwidth to their carriers, and a (large, probably) fraction of their providers' bandwidth to the long-haul carriers. (Realistically, Google is likely connected directly to the long-haul carriers.)

Given that, if my provider is unable to provide the contracted bandwidth at the contracted price, it needs to raise prices on new subscribers, and my prices when the contract is up for renewal. Google's provider has the same options.

In no case should my provider be charging Google for "their use" of my provider's network: it's my use of my provider's network, and I'm already paying for it.

Taking all that into account, to your first point, each provider charges their own customers for bandwidth, and uses the revenue thus generated to make improvements as needed.

To your second point, a user is granted a certain amount of bandwidth as a contractual term, and pays a certain amount of money as another contractual term. If your provider is having problems with customers using up too much bandwidth, their problem is with the overusing customer, or more likely, with their own network infrastructure. Getting you to blame your "abusive" neighbor for using up what he was contractually promised, rather than blaming your provider for building their network too small to meet their aggregate contractual obligations, is a good deal for your provider, but it's just blame-shifting. If your provider had made better usage estimates, or had built their network big enough to simultaneously provide all the bandwidth they were contracted to provide, there wouldn't be a problem. They didn't, and they're encouraging you to blame the wrong party for it.

To your third point, any provider who cannot charge sufficiently to build plant will go out of business, shrinking the number of providers and the available bandwidth, causing raised prices, making it profitable for the surviving (more efficient) providers to build plant. Basic market economics. The invisible hand, and all that.

So yes, bandwidth is finite. But it's also paid for. The free rider problem is not with the users, who are paying for the bandwidth, but with the providers, who want to charge people for services they are not providing.

Concrete example: for $500 per month, with no installation fees and no additional loop charges and a router included, I can get a T1 (1.5MBit symmetrical) to my house complete with up to a class C (256 IP addresses) upon providing justification for the IPs. Now, let's say that I wanted to become an ISP for some reason, but just providing connectivity. I could run wires to my neighbors' houses for a small cost; let's just say my four nearest neighbors so I could provide the connectivity wirelessly for a one-time cost of next to nothing (less than $1000, certainly).

Now, I could promise each of these four neighbors to provide them with a 740kBit connection, and could set up a DHCP server with the 256 IPs I would have registered, and the odds are that I would be OK. Unless all four were simultaneously on, and a couple of them were using massive amounts of bandwidth, no one would notice that I was promising more capacity than I have. But let's say that two of those neighbors are using Skype, and thus all of my bandwidth. Then you (a third neighbor) get on and find there's not much left for you. At that point, am I at fault for over promising bandwidth/underestimating usage, or are the two "free rider" neighbors at fault for using what I sold them?

Posted by: Jeff Medcalf [TypeKey Profile Page] at April 23, 2006 8:28 PM

I have no problem with any of your solutions, all of which boil down to the market determining the price, but you've violated one of the principles of net neutrality. What you describe -- the invisible hand -- is an x-tier network. The x-tier network is as big a bugaboo as app limitations, based on what I've read.

Now, you've presented, in your concrete example, a perfectly good mirror of how the internet has been built out to this point. What do you, the pipe builder and provider, do now? There's no incentive to invest, so far as I can tell, unless you go to x-tier pricing. Or metered usage.

As a business case, it's an interesting one. If I were SBC or Comcast, I'd just tell everyone that from this day forward, here's what you get (based on network metrics they have in hand), and if you want more, you're going to pay for it. Net neutrality from that perspective (a packet is a packet is a packet), but not from the other (x-tier). You simply can't have both.

Posted by: Scott Chaffin at April 23, 2006 10:59 PM

The invisible hand is not a characteristic of bandwidth shaping or tiered service levels; it is a function of how markets work, as described several hundred years ago now.

Let's start by disambiguating two different concepts: tiered pricing and bandwidth shaping. Tiered pricing is not related to network neutrality: it is the practice of offering more bandwidth/services for more money. For example, few people realize it, but DSL lines tend to be artificially constrained in bandwidth, so that providers can charge more for not restricting the lines. Bandwidth shaping is what violates the principles of net neutrality, by determining network costs based on characteristics of the traffic that is carried other than how much of it there is.

I don't see the argument to go to bandwidth shaping in order to cover costs. Let's say that we go back to my example, where I have 1.5MBit costing me $500 per month. (For the moment, ignoring multiple connections for redundancy and so on.)

So the most I could offer my four users would be 1.544MBit/4 = 386kBit. If I offer more than that, I could run out of bandwidth. In fact, I would actually have to offer them less than that, because there would be some bandwidth consumed by monitoring and other overhead traffic, so let's say (hypothetically) that I'm offering them 350kBit each, and not using tiered pricing (that is, I have only one service level). I have 4 users, and $500 in expense, so I would have to offer that at $125/user/month. I want to make a profit, though, so that I can invest in, say, upgraded bandwidth or advertising to get new customers or what have you, so let's make that $150/user/month.

That's not a good deal for the user, all things considered, when you can get 3MBit (shared, actual seems to be about one fourth that most of the time) with 5 static IPs for $110 or so here. But the principle still holds: I am offering the customer what I can reasonably offer for a price that covers that offering and allows me room to grow and take home a paycheck into the bargain. If the customer wants more theoretical bandwidth for less, he can go to a competitor offering bandwidth shaping or tiered pricing or making usage bets.

Now, I could offer more for less without bandwidth shaping by offering usage bets instead. And that is exactly what the providers have done, betting that average usage would be quite low and that even at peak times, they'd have enough bandwidth and IP addresses in their DHCP pool to cover all of their users. I have seen providers run out of IP addresses, and in that case they have no one to blame: they are the only ones that could have caused or prevented that problem. Similarly, they have caused, and could prevent, the bandwidth problem, at the cost of increased rates to the customer. Instead, they want to prevent the bandwidth problem by charging a third party, not part of the contract between themselves and their customer.

In my mind, that's just wrong. I don't have any problem with a provider doing that; I just would not use them. I don't feel any need to pay for their bad design decisions. On the other hand, I'm willing to pay more for good design decisions, which is why my connection is rock solid stable, gives me all the bandwidth I need, and provides me with static IPs — at about 2.5 times the cost of a vanilla DHCP or cable modem solution.

In the end, consumers will decide what flies and what does not. But I can guarantee you that no solution that requires a third party to pay for services is ever going to work. Companies like Google have less to lose than the local providers do: there will be competition against the local providers as soon as they start restricting access to popular web sites. And the local providers will change or lose business.

Posted by: Jeff Medcalf [TypeKey Profile Page] at April 23, 2006 11:38 PM

First, I understand the principle of the invisible hand. I know it's not a new description for bandwidth shaping.

Second, I've been designing networks for 25 years so I understand how they work, and more importantly to my business, where they break down and how to fix them.

Third, I'm not arguing for or against any position. I'm saying that there are pipe provider issues here that are being ignored. Skype itself is a great example of the usage bet, except in reverse. They know well that most nets are underutilized, so they can take advantage of that fact by selling SkypeIn & SkypeOut. Well, when you ramp up to 20 Skype users on a cable loop, all of a sudden the usage bet is a loser for two legs of the transaction. Skype starts failing for paying customers. The CableCo starts failing for ALL customers. You could go so far as to add the Skype user in there as a loser, if they're depending on it for business. So, in this case, who pays for the infrastructure upgrade? It has to happen if we want so-called net neutrality, as defined by all the utopians.

Fourth, yes, there is a part of the net neutrality push that is insisting on no tiered pricing. I'll see if I can find a link.

Fifth, I agree that we won't see third-party payer anytime soon. We will see the cost pushed to the consumer thru tiered pricing (it's the only model they know.) I've personally tried to work out how to take advantage of that as a competitor, but the services and technology aren't there yet (though they're tantalizingly close.)

Posted by: Scott Chaffin at April 24, 2006 8:53 AM

Here's the tiered pricing pushback article I read earlier this week:

I'm probably making a big assumption about the intent here, but it's what I read between the lines, knowing networks and knowing providers and knowing consumers. YMMV.

Posted by: Scott Chaffin at April 24, 2006 9:00 AM

Sorry; I did not mean to sound like I was talking down to you. Your sentence ("What you describe -- the invisible hand -- is an x-tier network.") was confusing, so I wasn't sure if you were confused or if I had not understood what you were saying.

I agree that there are pipe issues. My problem with what the providers are doing is that they caused the problem, and they are trying to fix it by paying third parties. Skype didn't cause the problem; while their software is horribly inefficient in terms of bandwidth use, they cannot use more bandwidth than the user has paid for. That the pipe is not built big enough is not Skype's responsibility. Nor is it Google's responsibility, even if large amounts of the data on the provider's network came from, or are going to, Google. Again, the failure to provide a sufficiently-sized pipe is the provider's problem, and theirs alone.

I understand why they did it. As I pointed out above, it's more efficient — if you (the provider) bet right — and thus cheaper. What is happening is that the usage patterns are changing, and the providers are skittish about raising their prices in accord with the requirements of their users. Again, this is not a broad problem: their users will go elsewhere until the situation equalizes again. The real solution, and I believe we'll see it soon, is either metered bandwidth by traffic, or higher prices for the promised bandwidth.

I am already paying a higher price to get certain assurances about bandwidth, so obviously it follows that I don't have a problem with this model. In fact, until this month, I lived in Keller, TX, where Verizon rolled out its FiOS (FTTH) service, and would have subscribed to it immediately if they had a service offering over fiber that provided static IPs. They did not, so I was not willing to subscribe at any cost. Had they provided static IPs, I would have been willing to pay more than I was already paying Verizon for a high-bandwidth business DSL with static IPs, because the available bandwidth would have been higher.

It's certainly possible that some people would argue both for net neutrality and no tiered pricing, all at what then become artificially low costs. It's a bad argument, and they'll lose it. The providers have to raise rates. What I object to is the providers' rent seeking, that is, going to the government to try to get them to force third parties to pay the providers.

That's wrong, and it's annoying.

I think we're pretty much in agreement, with two exceptions: I contend that the providers caused the problem (for justifiable reasons) and don't want to fix it by raising prices on their customers (for unjustifiable reasons), while you contend that a large part of the responsibility belongs to the creators or providers of applications and services that can take up a lot of bandwidth if it's available; and I contend that the person buying the bandwidth pays for the needed capital improvements, where you seem to be saying it's fine to push those costs off on other parties like Skype. If I'm misreading you, please correct me.

On the article that you linked, the first paragraph accurately sums up my problem with the providers: "Telephone companies in the US that provide broadband internet connections are starting to tread on dangerous ground. They are mooting charging companies that provide internet services such as telephony and video-downloading fees to ensure a speedy and high-quality connection." That is similar to charging Jaguar in order to raise the speed limits, so Jaguar owners can drive faster.

I think that you are misreading the article, or misreading me, though, on the idea of tiered pricing. When I say I do not object to tiered pricing, I mean that I do not object to the providers charging their customers for better service. When the article talks about a tiered internet, they are talking about the providers charging fees to companies that are not their customers. See the third paragraph of the article: "The effect of this discrimination would be to create a two-tier internet: users would find it easier and quicker to connect to services provided by the companies that paid such fees than others. Invisibly, customers would be steered towards these 'approved' services, including those provided by the phone companies themselves."

In the very next paragraph, the writer of the article agrees with me on tiered pricing: "There are a couple of obvious flaws in the phone companies' case. One is that they are already being paid by customers to provide the network connection: home users typically pay between $25 and $40 a month for a DSL or cable high-speed internet connection in the US. There is nothing to stop the companies charging more to customers who use the most bandwidth."

So in short, both the article's author and I seem to be saying the same thing: it's wrong to charge third parties for the providers' customers' usage, but fine to charge the customers for their usage.

Posted by: Jeff Medcalf [TypeKey Profile Page] at April 24, 2006 6:18 PM

Now you know as well as I do that the second a cableco or telco starts metering and charging the consumer for Skype usage, all hell is going to break loose. Of course, doing that is precisely how you charge customers who use the most bandwidth.

You're right, though, philosophically - the providers shouldn't be seeking relief thru govt. action. However, they are still regulated by the govt, so what do they do? Our leaders, elected by us, have trained them on how to get things done.

IMHO, the providers should be informing their customers of the true costs and capabilities of what they're paying for, and notifying them that their cost could go up or down. I shudder to think what we'll hear from the Lessigs of the world if they did that. If you ever thought the Digital Divide mewling is bad now, just wait.

Posted by: Scott Chaffin at April 25, 2006 8:23 PM

Information wants to be free. For wires, you pay cash.

Posted by: Jeff Medcalf [TypeKey Profile Page] at April 25, 2006 9:35 PM