Date: Thu, 27 Jun 2002 15:35:07 -0700
From: "Michel Py" <[EMAIL PROTECTED]>
Message-ID:
<[EMAIL PROTECTED]>
| but for the
| two different methods that ipv6mh is working on, there is no relation
| whatsoever between the network/fiber plant topology and the geographical
| aggregation.
I am waiting to see how this is going to be achieved. That is, all of
what you're promising in this message. So far, from what I have seen,
I don't see how it can be done. But maybe there's something out there
that I just don't understand yet.
| True enough. However, you forgot two important things:
|
| 1. If we make SLs with site IDs globally unique, there might be no
| assignement (or the assignement might be automatically granted by the
| addressing architecture, say a 37-bot hash of the MAC address, or
| whatever). With the geographic scheme we are designing, there would
| still be the need to go to a RIR and register a block, so we might still
| decide that someone with a dial-up 28.8k modem does not qualify for a
| portable, globally unique /48 block.
I certainly hope that isn't ever going to happen (with PI or PA
addresses). The aim is a /48 for everyone (who requests one), right?
What this has to do with anything I said I'm not sure though.
| 2. SLs do would not have a supporting protocol if used globally (how
| could there be one, if they are not supposed to be used that way). The
| way I see the geographical addresses is a support of the multihoming
| protocol. And if there is a protocol, there can be safeties and policy
| enforcement features built in.
Yes, there could be a protocol, with all kinds of things. And it can be
turned off, or replaced, if it gets in the way. Once RIP required that
there be only one subnet mask through a network ...
It isn't possible to enforce anything by relying upon protocols, we have to
depend upon agreements, and when appropriate "it is the only possible way".
The way we enforce aggregation now, isn't because there's some magic in
BGP that requires it, it is because we know of no other way to make routing
work. As long as that remains true, we don't have to create protocols to
enforce aggregation in IPv6 - it will just naturally be enforced, because
no-one can figure out how to not enforce it (nb: it doesn't really matter if
a few stray /48's or even /128's "pollute" the global tables - remember the
aim is to keep routing working, not to enforce the rule just because it is
a rule). If the assumption ("we know no way") ceases to be true, then
there's no reason at all that what we currently consider to be non-routable
addresses shouldn't appear in the routing tables. Ie: as soon as it is
possible to put them there, any policy aimed at keeping them out no longer
has a purpose, and should be scrapped.
kre
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
