Monday, June 12, 2006


Welcome to the Stupid Internet

At least Tom Giovanetti, of the Institute for Policy Innovation, a Bell-Funded Tank, writing in Friday's Mercury News, didn't call it a "dumb" network when he wrote the following science fiction . . .

Fast forward a few years to 2009 . . . Suddenly, the TV image goes pixilated, and then dark. The phone call drops. You hear yelling from your teenagers' rooms. But that's not all.

Across town, police on the beat suddenly can't reach headquarters on their radios. In an ambulance, the EMTs are trying to call in vital signs for a patient they are transporting to the hospital, but they can't get through.

Is it an alien invasion? A convergence of planets or some other astral phenomenon? No, it's a convergence of a different sort. Turns out that tonight is also the night of the Victoria's Secret Fashion Show, as well as the night Coldplay releases its latest song online. And YouTube has just released embarrassing video of a major Hollywood star having a ``wardrobe malfunction.'' Extremely high demand on the Internet is overwhelming available bandwidth, and regulations passed back in 2006 make it illegal for network operators to differentiate and prioritize content.

Welcome to the world of network neutrality, where all content is treated exactly the same . . . where somebody decided that a stupid network is better than a smart network.
Actually, Tom, I *noticed* that a stupid network was better than a smart network. I looked at "reality." I saw email, and the Web, and eCommerce, and Mapquest, and blogging, and Instant Messaging, and streaming audio on demand, and multiplayer online games, and many other miracles too numerous to list here, miracles that never arrived via "intelligent" networks.

Too bad Tom G didn't hear about what Internet2, the premier U.S. post-Bell-Labs network research effort found [.pdf]. (Research is how we learn about "reality.") Let's hum a few bars of Internet2:
“. . . our engineers started with the assumption that we should find technical ways of prioritizing certain kinds of bits . . . we seriously explored various “quality of service” schemes . . . [but] all of our research and practical experience supported the conclusion that it was far more cost effective to simply provide more bandwidth. With enough bandwidth in the network, there is no congestion and video bits do not need preferential treatment.”
In other words, when Tankers like Tom Giovanetti or Austrian-style economists like Alex Jacobson ask who decides how to allocate scarce bandwidth, they're asking the wrong question. Most of the cost of a network is obtaining right of way and constructing the poles, conduits, towers, antennas, cables, etc. Whether you provide a kilobit or a terabit is rounding error. Nobody needs to decide how to allocate scarce bandwidth. It is more expensive to allocate it than to simply provide more. Again, heed the research findings of Internet2 [.pdf]:
Simple is cheaper. Complex is costly.

In fact, Internet2 finds that GigE with reasonable service decisions is 10x more costly than simple GigE where the user decides what service to buy. GigE is enough bandwidth to run a telephone network for a city of 100,000 people, yet we have GigE interfaces on our computers. We could have GigE in our houses for little more than we now pay for 1000 times less capable services. Backbones are even cheaper than access networks.

Thus, the only reason to ask, "How do we allocate scarce bandwidth?" is if we're behind on technology. Even then, it is far more reasonable to ask, "How do we get ahead on technology?"

And if the telcos and cablecos won't get us ahead on technology, we should be asking, "How should we replace the telcos and cablecos?"

Technorati Tags: , ,

Enjoyed your entry on this. Profoundly disagree with you, but enjoyed your entry.

Will you admit you are wrong when all your nightmares and scaremongering don't come to fruition?
How will I know when I'm wrong?
This kind of reminds me of World of Ends. While I agree that stupid networks are better, this is a good argument against Net Neutrality legislation. If stupid networks are cheaper and more efficient, then do we really need legislation?
Great quotes from the scaremongering bellheads!

You might find my piece on net neutrality relevant as well. It has a pointer to a presentation I made about QoS in the context of net neutrality for Educause.

I read your article, and I read Mr. Bachula's Congressional testimony, and I noticed several things:

1) Your article contains a lot of hand-waving, and has very little relevant technical content.

2) Characterizing the issue as "smart" vs. "stupid" is deeply misleading. Complex vs. simple would have been far more accurate (and perhaps you already know that).

3) Given the blatant scaremongering in your own article, it's deeply hypocritical for you to accuse someone else of doing the same. ("Suddenly, the TV image goes pixilated, and then dark. The phone call drops. You hear yelling from your teenagers' rooms. But that's not all." Puh-leese.)

I was hoping your article might have something insightful to say from the opposing point of view, and I was deeply disappointed.

When people deliberately ignore empirical evidence and rely instead on scaremongering, it's usually because they believe their case would be weakened if their readers were better informed.
Nightmares and scaremongering.
- Tom Giovanetti

Hmm. I didn't see any of this in isen's article. I did see his observation that things have worked out pretty well, Internet-wise, so far. And I did see your prediction of blackening TVs.
I think the difference here is that the Internet2 folks had a very different kind of last mile problem. For the consumer market, there is a huge cost increase if you have to lay new cable, and phone companies in particular are getting close to the max they can do with the existing lines (sure, there may be some breakthrough, but for the moment this is where they are at). Once they have to replace those lines, the costs are MUCH higher than implementing some kind of QoS type solution. This isn't a case of QoS over gigE vs. 10gigE. This is a case of QoS over existing copper vs. installing fibre everywhere. It's not to hard to convince me that it's cheaper to install a million dollar switch than to dig up streets and front yards to get new lines to a few thousand homes.... and that ASSUMES that you can actually get approval for digging all that up (which might be tricky if only a hundred or so of those homes are currently broadband customers and only a few are currently asking for more bandwidth).
Post a Comment

This page is powered by Blogger. Isn't yours?