John M.R. Kneuer Acting Assistant Secretary Communications and Information gives a talk. He begins by talking about the value of the 700mH spectrum. It’s becoming available as “leapfrog” technologies are becoming available, he says. “What we really have is an opportunity for a game-changing opportunity against the first movers and incumbents.” [I think I garbled that, but so did he.]

“Net neutrality sounds very open, but it rapidly comes down to the government setting rate terms and rules for access.” “I firmly believe that market forces are going to provide an open network much much better than what we would get” through regulation. He refers to our current “great success.”

I ask the first question. I say something like: The great success has us at #19 in broadband access because there is no open market. I wonder what great innovation is going to come from the incumbents. We have proof that it doesn’t work because we’ve been trying it for about a decade [depending on how you count]. He says we’re asking for the government to set rates. I ask if anyone in this audience is asking for that.

Doc points out that wifi has succeeded because the spectrum was left open, not auctioned.

David Isenberg says that wifi wasn’t auctioned, and isn’t owned by a carrier, yet most people in this room agree that wifi is the most innovative sector in the entire spectrum. Kneuer agrees. David says there’s no business model, no carrier, and no market.

Kneuer: Wifi is local access to get to an underlying access. It doesn’t lend itself to building out broad networks.

David I: Same for last mile for fiber, DSL…these aren’t networks either. So they should be treated the same way…

K: They are nodes of a DSL network, etc. If I want to build out a 5mH wifi cloud, you won’t be able to scale it. Wifi’s authoriziation was for local area networks. It does not lend itself to the competing interests that need to be resolved in an efficient way. When you have lots of people trying to enter a commercial space and the gov’t is the bottleneck, the best way to handle it is in a transparent way by letting people bid for it.

KC Claffy from CAIDA kicks butt explaining how much bogus information there is — stats supporting the interests of the incumbents, based on bad research, without review or transparency. Fantastic presentation, but too fast for me to blog.

Now a panel on “Does the Internet need an upgrade?”

The first guy (I can’t tell who is who) says the Internet is us. The applications we’re using and the way we’re using them is radically changing. E.g., video vs. text. What we’re using it for now is different from what we originally designed it. And it will continue to change. The Net does need to expand and grow, but it’s up to each of us to determine how we’re going to effect that upgrade because we are the Internet.

Next guy says that when the End-to-End principle was created, every end point was trusted. Now we violate it all the time. So, we do need to look at the architecture, he says, if only to raise our collective consciousness. How do we get the balance right.

Van Jacobson from Cisco says the network has done pretty well. He says we’ve made only three architectural changes, yet it’s scaled incredibly. That’s because we’ve kept the network simple and moved the innovation to the edge. We solve the problems on the edges. We can do secure email by using encrypted messages. SSL, on the other hand, signs the envelope, not the message, which doesn’t guarantee very much,

David Isenberg agrees with 95% of what Van has said, but challenges his analogy. David says that all four panelists respect the end-to-end principle. The Internet grows, enables innovation, runs on any system, because of end-to-end. David wants to know why Van thinks security violates end-to-end.

Van says that SSL puts someone in the middle securing the envelopes, which violates end-to-end.

Van describes an attack where the man in the middle was fraudulent.

Isenberg: What’s the cost of fixing these genuine security issues? Everyone in this room knows there are security issues, but we still use the Internet. There are 40,000 traffic fatalities per year, but we treat it as a network externality. We still get into our little packets and get on our end-to-end highways…

Van: Vint Cerf said the middle can be arbitrarily untrustworthy, but we’ll fix it up on the ends. If the ends reject packets that aren’t answers to questions they asked, then the senders will learn to put enough in to let the recipients trust the packets. But that’s not our security model. Our security model is “Let’s make the center more secure.” That won’t work.

Isenberg: The incumbents would certainly say that the Net needs an upgrade because their business models are disrupted by the Internet. We need to beware their calls.

Q: Van, what about PKI?
A: It’s a disaster. It makes high-value targets.

Q: How do you keep the Net moving as people stream, etc.?
A: (Van) I started a company called PacketDesign a while ago that looked at the sort of data KC wants to see. We looked at the router downstream of the NBC Olympics coverage. It had 5,000 copies of the same data because the computer doesn’t know what’s in the packets. If we had a different model that saw the content propagating…As far as avoiding stutter goes, if you take every phone in the world and call up everyone in the world, it’d take 3TB, while a single fiber can carry 30TB. There’s room. There’s no incremental cost in adding more bandwidth. The incumbents act as if bandwidth is expensive as it was in the previous century.

[Tags: supernova2007 supernova07 commerce end-to-end net_neutrality ]

Categories: Uncategorized dw