[address-policy-wg] IPv6 PI for HOSTING

Le 25 déc. 2009 à 05:52, Masataka Ohta wrote :
> 
> As a person who changed IPv6 address structure from 10+6 to 8+8,
> trying to make IPv6 a little better than the worst, I know very
> well how IPv6 fails to work.

Not being such a person, I would be interested in learning in what "IPv6 fails", as you say.
(Just stating that IPv6 doesn't help to solve problems!)
Can you, please, provide a list of failures you identify for native IPv6 where it is available in addition to IPv4 (leaving out Teredo, 6to4 and ISATAP, which are not IPv6 per se).
References to written material would be appreciated.

I didn't work on 10+6 and 8+8, but my experience is that of a daily user of native IPv6.
As such, I reach in IPv6 various Google and IETF servers, which I use intensively, without having done anything special.
(I just enabled IPv6 in my OS which doesn't have IPv6 active by default).
Servers that advertise only IPv4 addresses are reached in IPv4 as from anywhere. 

Note that if more and more hosts would have native IPv6 (in addition to IPv4), and more and more servers would advertise IPv6 addresses (in addition to IPv4), then:
o less and less traffic would be IPv4
o NATs (still present for legacy hosts servers and/or service providers) would be less and less used and thus avoid port shortages
o peer-to-peer applications would work better where IPv4 is available only through cascades of NATs.

Regards,

RD