On 14-Jun-2007, at 14:09, james woodyatt wrote:
On Jun 14, 2007, at 02:56, JORDI PALET MARTINEZ wrote:
Just avoiding ANY collision risk. VERY VERY VERY LOW is not enough
for them.
My attitude is that IETF should tell them that's THEIR problem, not
OURS. Has the operator community explained why the odds of a
collision in a 2^40 address space pose an unacceptable risk to
them, which they can't mitigate without the participation of an
Internet Society organization?
If you're an operator, providing IPv6 access and presumably globally-
unique, globally-reachable address space to customers, why is it
necessary to use private address space to number your infrastructure?
You already have PI address space. Isn't it easier just to assign
some of that space for internal use (as the harmonised bits of the
RIR IPv6 allocation policy allows, I think) than to bother with ULA?
With IPv4, motivations of cost/scarcity might induce ISPs to use RFC
1918 addresses to number interfaces which don't require global
reachability. Those motivations surely don't exist with IPv6.
Perhaps by "operator" you meant something other than "ISP", Jordi?
Joe
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------