On Nov 4, 2003, at 12:48 AM, Tim Chown wrote:
On Mon, Nov 03, 2003 at 10:45:07PM -0800, Alain Durand wrote:
As I explain in a previous message, this last property is not verified by the hinden/haberman draft, as when those addresses leak, they would create untraceable problems, very similar to the one caused by RFC1918 leaks today.
But could we ever stop leakage?
And would it not be more dangerous if hijacked or randomly picked prefixes
leaked instead of well-known (probabilistically unique) prefixes?
You're assuming that the alternative to hinden/haberman is hijacking random prefixes.
I don't. I see allocation of real PI (by real I mean registered) a more serious alternative.
The more I think about it, the more I come to the conclusion that the issue
with the hinden/haberman draft is that those prefixes cannot be trace back,
making them as good (or as bad) as ambiguous.
I just saw a press release from a company building high speed network chips
that claim they can process up to a million route at 40 Gb/s...
so I'm honestly thinking that handing out PI to people who can justify the need
is not as scary as it sounds, at least it would enable us to wait until we get something
from Multi6.
- Alain.
-------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
