Bandwidth Cap Red Herrings

Time-Warner Austin has announced that later this year, it will implement a tiered set of bandwidth caps for its broadband customers. That's been a controversial and hotly debated proposal. I see some things frequently mentioned in the debate that I think are distracting, not helpful.

First, I see people frequently trying to draw a parallel between broadband tiers and cell phone usage plans. This makes sense only on a superficial level. Yes, in both cases you buy a certain amount and pay extra when you go over.

This analog does not hold up under scrutiny. The usage patterns of cell phones are pretty stable. There continues to be some migration away from landline service and towards online (VOIP) services, but for most of us our cell phone usage a year from now won't be that much different from what we do today.

This is completely untrue for broadband usage. New, innovative services are being created as we speak. Our broadband usage patterns may be radically different a year from now than they are today. Broadband caps lock in current usage patterns, which is detrimental to innovation.

Second, I see a lot of mention of average bandwidth utilization (6GB is the number quoted) and I don't think that's helpful. You really need to understand the distribution. You could have nine people doing 2GB a month and one person pulling 42GB a month. That averages to 6GB across all users, but that 6GB doesn't tell you what tier is appropriate for which user. (In this case, those 9 users would be happy with the lowest-level 5GB tier, and the other user would be a little screwed at the 40GB highest standard tier.)

The reason why distribution is important is because the Internet is widely viewed to operate on an early adopter model. If I'm an early adopter, I may be doing a lot of things with my Internet connection you currently aren't—but a year or two from now you may, as some of those applications may become mainstream and routine. Bandwidth tiers penalize early adopters and prevent new technologies from entering the mainstream. This, again, is a hit on innovation.

I'm sympathetic to the needs of broadband providers to manage their network resources. If, however, they cannot do so in a way that fosters innovation, then we've made a big mistake trusting them as custodians of our national broadband future.


Comments have been closed for this entry.

Thanks for posting this. I

Thanks for posting this. I know you've asked for people to comment at the Technology and Telecommunications Commission meeting tomorrow, do you think there will be anyone there?

Misleading points

While I appreciate the effort to cut through some of the BS around bandwidth usage caps, I don't think you've delved very deeply into the subject matter. Let's try to go a bit further.

Bandwidth usage caps are one of the tools that providers have to provide economical, fair-share access to the common facility we call the Internet. They address a fundamental error of omission in the Internet's architecture, namely the lack of per-user fairness. The Internet has relied, post Jacobson's Algorithm, on a per-stream fairness model that breaks down with the advent of P2P apps that use hundreds of streams per user.

Other mechanisms that have been deployed to address this problem are Fair Queuing (see John Nagle's RFCs about the early experience at Ford Aerospace with a global IP network in 1981,, rate limiting, and the QoS demotion system currently employed by Comcast. Caps are a very crude tool in comparison to these other systems, but they have the advantage of not requiring network upgrades to employ, and are effective across all versions of DOCSIS (or DSL, or fiber, etc.)

There is no law that says the cap of today must be the cap of tomorrow, so your fundamental line of argument is faulty. If the 40 GB cap proposed by TWC for your town proves to be disadvantageous, they can always increase it, and probably will. Comcast has doubled the rates of their basic and intermediate service tiers with no increase in price, for example. It happens.

Similarly, there is no law that says the light user of Internet access bandwidth must be required to subsidize the heavy user, any more than the light user of water or power must subsidize the heavy user (they frequently do on account of misguided volume discounts, but it's not fair.)

The legitimate aim of usage caps is to distribute fees according to usage, and that's a civic good. If they have secondary side-effects that harm innovation or some other social good, these can probably be addressed in the market, and only need to be addressed by regulation if there's sound evidence of market failure.

Wireless networks - 3G and 4G - have monthly usage caps as well, and they're typically set at 5GB/month and below. This is the proper analogy.

Similarly, while many of us have observed that "the bandwidth hog of today is the mainstream user of tomorrow," don't confuse heavy downloading with "innovation." The fact remains that most heavy downloading is porn and piracy, not the beta of Fedora 11. Genuine innovation in Internet services depends on the availability of low-latency service for things like Skype, and that too is adversely affected by heavy downloading.

In the name of fostering innovation, you're arguing for a policy that will actually hamper innovation. That's not where you want to go.

alternative to time-warner austin?

Is there a reliable broadband alternative to timewarner in Austin? I paid those bloodsuckers $175 last month and I'll be damned if I'm going to tolerate further gouging. They can dress it up in any "fairness" language they want, but I know the bottom line is they are looking for ways to dig deeper into my wallet. How long is Austin stuck with this monopoly? And where's my pitchfork?

And when we consider digital

And when we consider digital cable, internet, and phone bundles that TWC loves to sell, how much sense does it make for me to have my DVR pull down and record an uncompressed HD show, vs stream it from a website? If the argument was that heavy internet users are "hogging" all the network, it should be pay per gb, for tv and voip. Lets see them propose a cable TV pay-per-minute minute plan, and see who freaks out.

This is clearly, and shamefully, about profit. Would the poor innocent grandma that checks her email twice a week (whom heavy users should feel ashamed for hogging her internet) get a drastic price cut? I don't think so.