Sunday, August 20, 2006

Dreaming: The 100 to 1 Advantage

or
Intranet, Not Internet
or

How AOC and LUS Can Save The Internet

Ok, so maybe that last is a little over the top. But maybe not much over the top. Between them LUS and AOC could invent a model that would fix whats wrong with TV, make IPTV as currently conceived a bad joke, break the net neutrality debate wide open, and make locally produced content central to local video viewing. Sound too good to be true? A little like that other local wonder, Hadacol? Let me explain...

I've been ruminating over a recent post by Cringely that points to something I've mentioned before: that the costs of local bandwidth are very, very, small compared to the costs of interconnections. Your local intranet connection--in an office or at school has been running at 100 megs for at least a decade. Gig connections on an intranet are now common. But those sorts of speeds stop dead at the point where you have to buy internet access. For most of us around the US that means that those sorts of speeds just aren't real--almost of what we want to do is limited by, first, the availability of "big broadband" (you simply can't buy that sort of connectivity) or by cost (even where available it is prohibitively expensive for a single user to purchase.)

The choke point is in the last mile between your speedy intranet and the provider backbone that connects to the capacious internet backbone. An amazing number of our problems would go away if we could solve that last mile problem. Everything from Net Neutrality to international competitiveness to expensive and pokey internet connects are built around the last mile choke point and its monopoly/duopoly control.

The Suggestion

What Cringely suggests is that we avoid the choke point altogether. As a PBS employee he is concerned for PBS affiliate stations who are as worried about the death of the broadcast model as any other network station. Right now local stations are crucial in the redistribution of network content. The advent of cable has made a lot of non-local content available and reduced local stations' share of the market. They fear that the coming deluge of DT (downloadable TV) will make that worse, especially if their parent network makes the best content available from a centralized location on the net in a way that cuts them out of the revenue stream. That's not an unreasonable fear...Internet TV really might mark the end of local broadcast channels. Says Cringely:

But it doesn't have to be that way, because the supposed strengths of centralization aren't really strengths at all when viewed in terms of the much more imposing issue of bandwidth costs, where all the advantages are local.

I explained this to a group of PBS station managers meeting last month in Orlando, Florida. Where these folks tended to fear IPTV portended the disintermediation of local television, I argued the exact opposite. My reasoning came down to the price differential between Internet bandwidth and intranet bandwidth, the latter being that bandwidth entirely within the ISPs local point of presence or data center. There is a lot more of this intranet bandwidth, for one thing. Depending on how their network is segmented, a local provider of cable Internet or DSL service may have gigabits of aggregate customer bandwidth attached to a much smaller Internet pipe. A 100-to-one ratio of internal to external bandwith is typical, meaning the effective cost of internal bandwidth is 100 times lower.

What I advised the station general managers to do was to serve their traditional audiences as much as possible over internal, rather than external, connections. This means colocating a server down at the telephone and cable TV companies, which isn't hard to do since most communities have just two broadband providers, and the PBS station manager probably knows both of them from Rotary meetings or from the local United Way board.

Cringely tells PBS affiliates to locate a nice healthy server inside each ISP intranet (like Cox's or BellSouth's) using their Rotary ties with local telecom execs and then to push out the content they own as downloadable video. Content pushed inside the local Intranet would get the larger pictures faster intranet speeds would buy it and those outside the intranet would still get the postage stamp versions we've grown used to. Downstream the dream is to cut revenue sharing deals for good national content and evenutally build up a network of non-profit video servers based on the new PBS infrastructure. At that point the new non-profit infrastructure would be positioned to schlep bits for any commercial provider at its non-profit rate. The combination of better quality and cheaper carriage would make the new network a natural winner.

A pretty picture.

But one that overlooks a lot of reasons why it won't work as advertised.

Cringley's Mistaken About Curent Local Networks:

First and foremost the self-interested nature of the incumbents will ensure that they won't just give or even sell Public Broadcasting the rope with which to hang them. The death of the broadcast model would turn the current cable video providers into sellers of commodity bandwidth. They hate the idea to their very core and the current net neutrality fight is largely based on the telephone companies wanting to avoid becoming commodity bandwidth providers. Giving away the farm to local PBS friends would be a policy decision of the first water and the guys PBS managers schmooze with at the local chamber of commerce aren't the ones making those decisions.

On top of that, unless I very badly misunderstand how local networks are set up the idea that it will be easy to let every user make use of excess "intranet" bandwidth is very mistaken. For the Bells the extra bandwidth simply is not there. If they had it they wouldn't need to be messing around with IPTV and trying to squeeze enough bandwidth for a few switched TV channels out of their next upgrades to the system. (They can't push any video now.) The cable companies do have more bandwidth streaming past your door than they give you--much, much more, enough to since they are streaming many full-screen channels of video past your connection at every moment. But they have no facility to differentiate between IP address inside and outside the system--and the speed cap they set on your internet access is controlled in your cable modem. My sense is that they'd have to seriously re-engineer their systems to give you the unbridled access to the local network that would make Cringely's plan work. And as I've pointed out already, they've got no reason to go to considerable expense just to enable the destruction of their business model.

So, while Cringely's dream is very suggestive, and points to the real potentiality of an intranet, it doesn't seem at all likely to be realized unless some outside force pushes it on the incumbents as a matter of competitive survival.

But Cringley's Basic Insight is Correct:

Here's where AOC and LUS come in. (You anticipated that, didn't you?)

The main reason that Cox and the cable companies won't participate is self-interest. This plan would be great for customers who'd get internet service fast enough and cheap enough to replace the old broadcast model of the internet. But it is not the way to maximize their profits, shareholder return, and executive pay packages. Even if it would maximize the value returned to customers for their monthly investments.

What you need for Cringely's vision to work is a network owner who isn't in it to maximize profits for distant shareholders or increase executive bonuses. A network owned by its customers. A LUS. LUS has already committed to engineering their system so that in-system bandwidth will be fully available between its local subscribers. (Langiappe consequences of 100 megs: Telework, Grid, Futuristic Consumerism) The initial thinking was that the commitment would yield a 100 megs of internal speed on local subscriber interconnections. By the time the system is built that will, more likely, be a gig. That is much more than enough to offer full-screen HD in multiple streams to each household in the city. LUS would have bandwidth to burn in the last mile and little reason not to give its customer-owners exactly what the desired on the internal highway.

LUS would make a rational ally for AOC in moving to a local, server-based version of its current programming.

Natural Allies:

There'd be a lot of advantages for both AOC and LUS. All of AOC's shows would effectively be playing in "prime time." There'd be no necessity to try and decide who goes on when for most of their shows. There'd be an easy way to gauge demand for their shows.

The local producers would love it that their shows would not vanish as soon as the red camera light goes off. With smaller downloads available outsystem local creative types could build an audience outside Lafayette that would help develop an ecosystem for future growth of their shows. (Austin city limits is a model that could work well in Acadiana; a genuine Cajun or Creole food show could also be successful nationally after building an online audience. Current TV models suffer from not having a genuine "farm" system--all shows originate nationally and local sensibilities--and creativity--have little influence witht the "coastal" executives.)

For the customer this type of system would have all the advantages of a huge, online DVR/personal archive of AOC video. It would be a free, permanent online library of locally produced content. (If you wanted to go back and see that contentious (cough, cough) council meeting that everyone still talks about you could.)

LUS would love it because they'd have a huge and clearly understandable differentiator of product. They'd be able to have content on their system in a form that would be impossible for any other network provider to match. There's no reason the online version of AOC shows couldn't be provided in an HDTV alternative. HDTV steaming video over the internet side of local connections would be easy for LUS's fast internal connections to handle. But BellSouth/AT&T's and Cox's networks simply could not handle that sort of load without an upgrade of a sort they are not even beginning to contemplate. Installing an upgrade of a sort which would allow one stream of 25 megs HDTV--much less reengineering the system to provide that much bandwidth between any set of local users regardless of the tier they've bought-- is not on the map of either BellSouth or Cox. On the other hand, LUS is planning to build just such a system; one that would support at least 3-4 independent HDTV streams to each household. By supporting AOC in this way LUS would be making it clear that its system could support services that the other networks could not--and would be expressing its mandate to encourage local alternatives.

Okay, and PBS too:

With the AOC server up and running on the local LUS network Louisiana Public Broadcasting (LPB) would have a local network provider whose interests are in align with their mandate to serve the people of the state. Cringley's vision could be realized in Lafayette--even if they'd have to talk to Ed Bowie over at AOC instead of the cable reps at the local Rotary. As a social studies teacher in remission I can tell you that LPB owns a fantastic library of video. It's mostly inaccessible to the public that paid for it. But there's no reason, save cost, that it can't be online. In Lafayette that cost could be reduced to the price of space on AOC's server farm. A collegial agreement could surely be reached since all the participating parties would benefit from the exposure. (Video served outside the LUS region could go out in the severely reduced postage stamp format we use now and be paid for in at-cost bandwidth or a trade for making the content available if the right deal were cut with AOC and LUS.) Maybe LPB could begin recruiting membership from all over the country--like KBON (stream) and KRVS (stream) do now. LPB already has some downloadable video; for instance it offers the news magazine "Louisiana: the state we're in" online. Why not get muscular about it? The grant to digitize all their old shows should be easy to get.

And there's absolutely no reason to be selective about this...LPB (and AOC, for that matter) could offer their content to be placed on BellSouth and Cox local servers just as Cringley originally suggested. Just don't expect those entities to be willing, or able, to take up the offer.

And that's only the beginning--there are lots of archival video and voice recordings whose collection protocols mandated that they be made publicly available languishing in obscure corners of university libraries. Both ULL and LSU have astonishing oral history archives, for instance. Every heard of them? Thought not. While the Library of Congress has done fantastic job of making those sorts of things available local content gets the short shrift. An initiative like the one described here could change all that. I know some folks at ULL are interested...

Back to the outrageous thesis—LUS and AOC save the Internet:

Right now we experience the bottleneck that is at the root of this discussion most painfully when we try to download those short postage-size video clips from the web that were pioneered by news sites but have morphed into a major cultural phenomena by the rise of sites like YouTube. The incumbent providers claim that they need new revenue sources to bring you up to speed (a bogus claim) and want to extort a surcharge on top of what companies like Google, Vonage, and YouTube already pay in order to guarantee them adequate service over that last mile into your home.

That last mile is the only place where the current incumbents have enough monopoly control to impose such an unpopular and unfair regime. Take away that control by offering the sort of robust local competition that would be offered by an LUS fiber to the home project which would have no problem with commoditizing its bandwidth if it served its owner/customers and the net neutrality issue would vanish. Without their current duopoly control over the last mile the incumbent providers would have no leverage to impose the "packet prioritization" that is at the heart of the Net Neutrality debate.

What's more there'd be no need. With sufficient bandwidth there is no need to worry about packet prioritization and, in fact, the additional overhead would just drag down the network to no useful end. (This is the conclusion that the Internet2 folks reached after trying to improve an advanced broadband network by packet prioritization. They decided it was easier, cheaper, and faster to achieve the goal of new internet services by simply increasing the bandwidth.)

It's not just the hot net neutrality issue that is built around the current choke point between intranets and the internet. So is the continued survival of the old "broadcast network" model of TV. You know the one that forces you to watch fixed length shows, interspersed with truly stupid ads, at fixed times. (Die TV! Die, Die, Die!)

So maybe LUS and AOC could provide the "demonstration proof" that big local bandwidth, locally owned, and operated in the interests of local owners could save the internet and, as a bonus, kill off the network/cable model of video production and consumption. Not only would our local community love it but it would serve as an example to other communities across the country.

It's not really all that far-fetched. It just requires a bit of vision and some modest courage and cooperation.

Why not?


(Dreaming dept.)

No comments: