Thursday, December 15, 2005

Your Big Bandwidth Future


Why you do SO need a 20 meg line into your home.

One of the more annoying aspects of the seemingly endless fight against the incumbent telecom providers has been the tiresome repetition of some version of "you don't need this." I've even heard it repeated by consultants who really should know better. (This version goes something like "What can anyone do with 20 megs?")

The short and easy response is that such noise is best ignored. Anyone who's been a telecom user for a while recalls that time when v90 dialup modems supplied more bandwidth than you needed for any html page on the web. (MP3? real motion? Flash? None of that existed; this was the age in which the blink tag was the foundation of fancy graphics) "What do you want with more than 56K?" was actually asked...especially if you (gasp) had to pay a special fee. People asked the same thing about why anyone would want a second telephone. Or a wireless handset. Or a mobile phone.

That sort of noise is best understood as the "Argument from Lack of Imagination." It's a bad habit we should have gotten over sometime around 1870 when the pace of technological innovation picked up. ("Cars, who needs cars, smelly, expensive and unable to reproduce!")

While history teaches us that such cavails are reliably unreliable we've hit a spot where the nature of the change in front of our generation is beginning to emerge from the's possible to at least point at the use for 20 megs.

The biggest picture answer is: EVERYTHING.

Right now you're hooked up to different, specialty "service" networks each hand-crafted for the particular narrow service it sells you. Phone, Cable, Cellular. All that is in the process of vanishing. It is being replaced by a cheap commodity: digitized information carriage. Specialty transport systems are inefficient and expensive by their nature. --If you needed different roads to transport clothing and food and neither could be used for anything but clothing or food your clothing costs would have to pay for the "clothing network." Only the fact that we have an uspecialized commodified transportation network which spreads the cost over many products and uses allows transportation of clothing to be a fraction of its cost rather than the dominant element. The expense of constructing such single-purpose supply systems in telecommunications is part of what makes each such system a natural monopoly--an economic fact which raises the costs still more.

Where single-purpose networks dominate it is hard to introduce new services since each new "product" would have build a new network to provide it. (This is the real reason we don't have universal video phones and video conferencing: since the old specialty phone system won't handle that service the phone companies would have had to build entirely new, even more expensive, specialty systems which would cannibalize their paid-off system.)

Currently the Internet service we get is tacked onto--in truly crippled ways--the old phone and cable systems. They weren't designed to offer internet service and they provide it poorly, in ways that radically limit capacity (which translates for the user into limited and unreliable speed). Even so, the internet service we have now is an instructive example of how destructive unspecialized networks can be. Designed to be unspecialized, designed for the unimpeded flow of information, built around a densely interconnected network with no central locus of control the internet conceptually resembles the interconnected mesh of our road system. Even running on networks that cripple its power it makes the flow of information a commodity not a specialty product. No bit costs more than any other bit. If you've got the bandwidth you can talk and video conference for no additional cost.

This is what the owners of specialty networks truly fear: a real internet network; one that commoditizes the flow of information and, inevitably out competes their lucrative, narrow monopolies.

A network on which EVERYTHING will run.

That's what the future holds. And Lafayette will be getting it early. The modern network, optimized for commodity information, for bits, is already in place in the backbone of all of today's networks. It's based on fiber, communication protocols that mostly don't care what the bits describe, and it is characterized by capacity to burn. The final refuge of the limited capacity, specialized network is in the run to your home. Replace that with big bandwidth fiber to match the backbone and the floodgates will open. Even with the limited bandwidth now available you can see both voice and video trying to squeeze itself into the crippled internet portion of the networks that enter your home. (VOIP in all flavors and every video you ever watched on the net are evidence of this.)

That "20 megs" will be carrying EVERYTHING, not just a slightly souped-up internet. Cable TV will die (die die), all voice will move to the internet and the owners of last mile fiber will surely leverage their excess bandwidth to provide a (quadruple play) WiFi/WiMax system that will vastly reduce the need for specialty cellular systems. You'll download your shows and drop cable--and if you think you watch to much TV when you don't like about half of what you actually watch consider how much you might watch if everything ever made was available, instantly, for a modest price. (We'll need to develop some self-discipline!) Talk as long as you like to anyone you like anywhere in the world. Talk, hell, videophone them. Better yet, video conference on the fly. (Trying to get all your siblings to tell you what the nieces and nephews want for Christmas (and what their parents think they should and shouldn't have)? Knit it together with a video conference, or just use it to settle on a date for the granpop's birthday.)

A couple of video conference streams plus an HDTV stream of Barney to distract the kids at your house while you play Santa could chew up considerably more than 20 megs. Its dead easy to think of scenarios where a 20 megs limit could be hit pretty easily. It doesn't take all that much imagination.

There is more, of course: the specialized nature of current networks means they don't interact very well and, as already noted, erect barriers to entry for new services. Modern networks won't have that problem it will be easy start new services and easy to link them. Open a window in your video conference for a froogle search, someone else could be simultaneously calling their "Toys Are Us" for local availability. Coming up with all the new services and interconnects will take imagination. But I am sure that some enterprising entrepreneurs will prove up to it.

The point is that we'll find it easy to chew up bandwidth. Most obviously through generous use of video but also by loading multiple, interacting streams of information. And since it will all be coming in through that one, modern, commodity pipe we'll be getting it cheap and fast. That old 56k modem seemed nifty at the time and so will 20 megs. For about the same length of time.

The fine print:
Sure, sure there is a caveat here, there always is...the commodity pipe that I describe is a text-book description of a natural monopoly. Expensive infrastructure (the "pipes") and cheap, commodity "product" (the digital "bits") make for a classic utility situation. Just as with the provision of water, providing a universal, expensive infrastructure to all to sell a cheap product means that there will be only one provider...a second provider who split the number of subscribers with the first provider would just force everyone to prices which would necessary to pay for two sets of inrfrastructure when one would do easily an yield no better service (water is water, bits, bits). In the end only the cheaper provider would remain. In the end there will be only one. A monopoly. This is not a welcome outcome. But monopoly is the only realistic possibility for a converged big pipe network. Given the stone-cold fact of the matter I'm glad Lafayette's last mile will be controlled locally and publicly. A utility or a coop is by far the best solution.

Really small print: I think a lot of good-hearted people are having a hard time facing the cold truth of an system evolving pretty inexorably toward monoply. For folks without a local public utility or coop this is going to end up badly (IMHO). Locals can insulate themselves, at least partially from what is coming by doing what Lafayette is doing--building their own utility to free the last mile from private monopolization. The real limits and the real monopoly is in the last mile...alternative backbones are already readily available. But for our nation as a whole the answer will have to be federal and aimed honestly at regulatory control or public ownership of what is clearly going to emerge as the 21st century's central monopoly utility. We did a mixed job with electricity in the 20th. We are doing a lousy job with communications networking right now. We won't do better until we buck up and find the courage to admit the situation we are in...and the need for a willingness to face hard truths applies, especially, to those who believe they are fighting for consumers and communities.


Peter M. Allen said...

Well spoken Scott! As the concept as an Information Utility matures, it will be driven by the certainty that "the taxi company must not own the road". If UPS owned Main Street, how could the USPS, FedEx, Emory, or DHL compete? Network ownership and operations need to be separated from content/service creation and delivery - and fair access needs to be a mandate.

A First Mile Optical LAN would replicate for residents the same corporate infrastructure without which we cannot do our daytime jobs. An information utility would keep a lot of money and traffic local while providing competition for services providers in and out of the utility. America is about liberte' - which assumes having a choice (ergo competition).

To say that we'll never need more than 20 Mbps to our homes ignores history. Look at the development of our telephone and electric industries in the last century. And remember also: "We'll never need more than 640k" of memory in our PCs.

We are at a flux point in our country's Internet infrastructure. We are way behind many countries in broadband deployment. I live in Silicon Valley and have worked at Napster, Bell Labs, and eBay. I have optical fiber to my home compliments of our local electric utility, but our four year long FTTH trial unfortunately was concluded and went dark today [sigh]. I know the risks of saying "There's never time to do it right, but always time to do it over". I applaud y'all in Lafayette for knowing and doing what is right!

John said...

Hi Peter,

I presume you live in the right neighborhood in Palo Alto? I was sorry to hear they got cold feet. It sounds as though the experiment was a success but the city couldn't quite bring itself to follow through.

It must hurt to go back...

Oh yeah, not sure who Scott is but he's not the author of this post. :-)

Peter said...

Doh! Thanks John. Well written! Yes I live in the former Palo Alto FTTH Trial area. I have both cable and DSL balanced now into my home to make up for the speed and reliability I had with FTTH. Sigh.