On Wed, 26 Mar 2003, Tony Hain wrote:
The only relationship SL has to renumbering is the ability to have
connections persist while a network is intermittently attached to the
public network. Renumbering is already solved in terms of the simplicity
of moving hosts from one address space to another. The complex issues to
work on are the places like firewall & router configurations that have
explicit addresses in them. What is not fixable is the fact that apps
will break if you change an address out from under them. This is a fact
the app developers complaining about the complexity of scoped addresses
continually overlook. The assertion is that all a network needs to do is
change the addresses in use when connecting. Never mind that every local
use app will break on every one of those events. That is not an
acceptable approach.
Who said the addresses are *completely* revokated when the network
connectivity is intermittent?
More likely than not, those address advertisements have a lifetime longer
than the duration of the downtime (both preferred and valid in RFC2461
terms!) -- and whoops, everything works like a charm still!
-----Original Message-----
From: owner-ietf(_at_)ietf(_dot_)org
[mailto:owner-ietf(_at_)ietf(_dot_)org] On
Behalf Of Eliot Lear
Sent: Wednesday, March 26, 2003 12:59 PM
To: alh-ietf(_at_)tndh(_dot_)net
Cc: 'The IETF'
Subject: Re: Fw: Welcome to the InterNAT...
Tony Hain wrote:
Trying to use SL for routing between sites is what is broken.
But that's not all...
The space
identified in RFC 1918 was set aside because people were taking
whatever addresses they could find in documentation.
Not as I recall. Jon Postel received several requests for
extraordinarily large chunks of address space, particularly
from Europe.
I believe Daniel Karrenberg might have more information.
This forced
his hand. In addition, people such as Paul Vixie were trying
to do the
best they could to make random address space sork, which is
admittedly a
trick in a small name space. Recall at the time that CIDR was a new
thing. You couldn't simply use a portion of network 10, for
instance.
The same cannot be said for IPv6.
SL was set aside because
there are people that either want unrouted space, or don't want to
continuously pay a registry to use a disconnected network.
Any address space can be unrouted address space. Fix the underlying
problem, Tony. Making renumbering easy. If we don't do
that, IPv6 is
no better than Ipv4 (with the possible exception of MIPv6).
It is far
cheaper to train an app developer (though there may be an
exception or
two) to deal with it than it is to fix all the ad-hoc
solutions that
people will come up with to replace SL.
Fix the renumbering problem and this isn't an issue.
Eliot
--
Pekka Savola "You each name yourselves king, yet the
Netcore Oy kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings