[e2e] Re: crippled Internet

Vernon Schryver vjs at calcite.rhyolite.com
Wed Apr 18 11:37:07 PDT 2001

> From: RJ Atkinson <rja at inet.org>

> ...
> >If @Home only wanted to limit
> >bandwidth, then it could by saying so in its AUP.  
>         Not according to the lawyers -- according to the lawyers,
> to do a pure bandwidth approach would require @Home to provide
> authenticatible usage data for each subscriber -- which is not
> feasible with currently available commercial cable modem
> equipment.  We were told this was based on precedents in case law.

I translate that as "If @Home were to charge for bandwidth, then @Home
might have to provide some minimum bandwidth, and heaven forfend!"

I can't translate "authenticatible usage data for each subscriber -- which
is not feasible with currently available commercial cable modem equipment."
No one (except perhaps a compliant lawyer) can honestly claim that either
the subscriber equipment or the head end equipment needs to have any more
authentication machinery than it already has.  It's not as if the cable
outfits are not using DHCP, or that DHCP does not involve matching IP
addresses with a sufficiently authenticated subscriber token such as an
IEEE MAC addresses (PPPoE "Ethernet" address).  Forging MAC or IP
addresses would be self-defeating for most (all?) end-to-end services.
It's also not as if bandwidth throttling or even accounting couldn't be
done upstream of the cable head ends, and that more than blindly throttling
bits per IP address regardless of current IP address user would be required
if @Home were concerned about bandwidth instead of "servers."

Am I confused, or did @Home announce about 6 months ago that it was
installing just such automatic bandwidth throttles?
Judging from http://news.cnet.com/news/0-1004-200-344209.html?tag=bplst
and http://news.cnet.com/news/0-1004-200-1463562.html?tag=bplst
I am confused--about the date.

] From: RJ Atkinson <rja at inet.org>

] 	Your apparent experience with the legal system
] is significantly different than that of (most) ISPs.
] In any event, this is off topic for E2E at this point,
] hence mute.

"mute" or moot"?

Off topic because it contradicts your claims about the legal system
we both enjoy?

That major ISP's iincluding @Home are trying to end the existence of
the end-to-end Internet sounds somewhat relevant to an IRTF mailing list
about end-to-end services and related issues.  If they succeed, then
this mailing list itself would be moot.  Granted, it's not about
ECN bits.

It's not only subscriber-server part of end-to-end services and access to
arbitrary remote SMTP servers that are under attack.  There are many
reports of port-80 HTTP redirection proxies.  So far I've not heard of
the obvious next step analogous to adding advertisements to traffic forced
through port-25 servers by port 25 filters, but it will happen.  (I assume
that Netzero's forced-advertising mechanisms involve techniques other than
the obvious HTTP hacks.)

} From: "Henning G. Schulzrinne" <hgs at cs.columbia.edu>

} At least in the rural United States, party lines were in use well into
} the 1950s. ..

} As an example, http://www.coloradovalley.com/cvtchistory.html says "By
} the early 1970's, all of the cooperative's members were converted to
} single-party service.", thus, my time estimate may be somewhat
} optimistic.

I think that multi-party lines were all that were available in 1980's and
I think the early 1990's in parts of the the Colorado Front Range served
by U.S.West from Coal Creek Canyon north toward Estes Park, all north west
of Denver and partly in the Denver metro. calling area.

> From: "Andrew Smith" <ah_smith at pacbell.net>

> Of course there are other practical issues, involving private and/or dynamic
> IP addresses and PPP-over-Ethernet tunnels (yuck!) that are typically forced
> on you as a residential DSL consumer these days, which make it hard to enjoy
> jon's Real IP (tm). It could be a long wait for those TCP connections to
> arrive if nobody knows today's IP address. Some simple web page
> "indirector", updated every time the server IP address changes, could
> provide the DNS function I suppose. Methinks we need better FOGLAMPS.

Ask Google about dynamic DNS as in http://www.google.com/search?q=dynamic+dns
Many people claim that it works just fine, although I've always
had trouble believing in its reliability.  On the other hand, people
I know with cable modem service report that their so called dynamic
IP addresses don't change for years at a time.

Vernon Schryver    vjs at rhyolite.com

More information about the end2end-interest mailing list