NNSquad - Network Neutrality Squad

NNSquad Home Page

NNSquad Mailing List Information

 


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ NNSquad ] Re: "Regulating" the Internet -- and Distinctions -- ATM vs IP


Decoupling the physical facilities from the services is the pragmatic and sustainable face of neutrality. This way we avoid the conflict of interest that I consider the defining property of today’s telecom. Too bad the FCC is limited to playing word games as they dance around this problem. I remember watching Bob Pepper trying to say this within the constraints he faced. Yet there people in the audience who couldn’t get it and seemed unable to recognize this and tried to act as if the rules made sense.

 

There is a viable business in maintaining the transport if we fund it as such.

 

"How robust is ATM" is a key question. This is why protocols do matter. ATM is good for links but doesn’t composite well. The genius of the end-to-end argument is in reducing the role of the network to best-efforts packets thus shifting the design point from one in which any failure is fatal to one in which anything path that works is success.

 

This is also the source of conflict – telecom is all about the path and the Internet is about what we can do independent of the path. They are very different views of the world and the source of our problems is in hobbling the benefits of path independences with all the limitations of path dependence.


But today’s IP protocols are a mixed bag because the IP address reflects the confusion of the path with edge-to-edge relationships. Once we decouple the use of the IP address as a path from its role as a IP address as a path from it's at we can do independent o with the curent oren end-point identifier we are ready to take on telecom. We no longer need to petition for an IP address and thus be beholden to service provider. And routing is simpler because the path identifiers can be used for network management.

 

We need to decouple the roles of the IP address as well as the elements of telecom if we are to get the benefits of neutrality.

 

We can then have a viable business is installing and maintaining the physical facilities and pay for it directly rather than having to maintain scarcity in order to fund it out of service revenue.

 

-----Original Message-----
From: nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org [mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On Behalf Of Barry Gold
Sent: Thursday, September 11, 2008 12:38
Cc: nnsquad@nnsquad.org
Subject: [ NNSquad ] Re: "Regulating" the Internet -- and Distinctions

 

Phil Karn wrote:

> What has to happen is very clear: "telecommunications services" must 

> be defined as "physical transmission facilities" and "information 

> services" as the applications of those facilities. The former must be 

> carefully regulated wherever a market failure has created a 

> destructive monopoly. The latter should be left completely alone 

> because it's vibrant and diverse, and there's really nothing wrong 

> with it.

>

> Under no circumstances must the telecommunications service providers 

> be allowed to grab further control over the information services that 

> use them. We can count on them to try.

 

I think you have found the exact crucial distinction.

 

Much as I hate the idea of telling the people who built out the physical

lines (cables, fiber) that "we are taking control of your stuff for our

benefit", it worked out not-too-badly in the telco "breakup".  So maybe

an equivalent thing here: a requirement that the cablecos rent the use

of their lines to other providers at a price determined to be "fair" by

the government.

 

That would allow almost anybody to get into the ISP business for the

cost of a couple of servers and a few routers, just as you could become

a "telephone company" for a few $1000 in the 1980s & 1990s.  Of course,

you won't last long unless you can sell your service for more than the

cost of renting (your share of) the cables plus your own operating costs...

 

Another open question: how robust is ATM?  If we allow random people to

declare themselves ISPs and connect directly into the ATM "backbone",

will that create the same sort of security holes (at a much deeper

level) that we currently see on the Internet, with DNS Spoofing and fake

routing tables?