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

Re: Accept hain/templin draft as wg item?

Subject: Re: Accept hain/templin draft as wg item? Date: Mon, Aug 25, 2003 at 10:39:01AM -0400 Quoting Keith Moore (moore@cs.utk.edu):
> I guess that's the core of the disagreement.  I don't see our task as that
> of replacing SL - certainly not in terms of coming up with "a" replacement.
> I recognize that SL was designed to fill certain needs - or probably closer
> to the truth is that lots of folks were projecting needs onto SL.  Where
> there are legitimate needs, we need to try to address them.  It's also
> clear that PA addressing alone cannot meet some legitimate needs.  What's not
> clear is that there is a legitimate need for "local" addressing - and stating
> things in that way seems to presume a particular kind of solution that we
> already know to be problematic.

This is close to my perception -- which is why I somewhat harshly advocated 
for the document to be caught and non-sent to IESG. My apologies for choice
of words. 

In this situation I am pretty convinced that we have to deal with
varying cost tradeoffs more than finding Graals of address supply
management. I think that the price paid for "wasting" global address
space by filtering it off at the IGP edge in order to create
non-routed (as opposed to non-routable) address space is worth

Having recently been provoked in bit-flipping directions by external
input I do see a need for stating *why* I oppose creating dedicatedly
local addresses - it is simply going to occur crashes between mirror
worlds. The discussions on LL address usage are typical for this
-- the LL scope was created to cater to specific functions (RA et.
al.) but the distinction is blurred as soon as the notion "an address
is an address is an address and as long it is unique I can use it
pretty much as I please." takes over peoples minds. It sure has
taken over my brain and it is quite contagious, on both sides of
this long debate. Thus, there is an operational requirement to remove
potential sources of ambiguity because the usage patterns for
addresses tend to approach a state where every service may be
deployed on any address. P2P certainly is a prime example.

Good arguments have been made that this is simply a failure to release
v4 paradigms, and I'm tempted to agree -- but that does not mean
we have a solution to this except forcing uniqe address space onto users 
and programmers.
Måns Nilsson         Systems Specialist
+46 70 681 7204         KTHNOC

Are the STEWED PRUNES still in the HAIR DRYER?

Attachment: pgp00059.pgp
Description: PGP signature