Re: [v6ops] 464XLAT

2012-09-06 15:16, Cameron Byrne :
> 
> On Sep 6, 2012 1:03 AM, "Lorenzo Colitti" <lorenzo@google.com> wrote:
> >
> > On Thu, Sep 6, 2012 at 4:54 PM, Maoke <fibrib@gmail.com> wrote:
> >>
> >> how can we ensure the above addresses do not conflict with any manually, arbitrarily configured addresses? i am lost. :P - maoke 
> >
> >
> > AIUI, we say "conflicts are impossible, because this is a reserved range", and then we cross our fingers.
> >
> 
> I agree with Lorenzo's sarcasm.
> 
This "sarcasm" isn't deserved: one who manually configures a host with an invalid address with regard to RFC 4291 should cross his fingers in all cases, with 464XLAT or not.

Besides, it is particularly unexpected from you, after you stated "In the case of the Android CLAT code that was submitted to AOSP, the /64 is simply copied from the WAN 3G/4G interfaces to the LAN. No magic. No ND Proxy. No DHCP.  But, it works for all known cases".
With this, collisions between CLATs and host addresses are possible even with well-behaved hosts => The need to cross fingers is permanent ;-).
By using the proposed IANA-EUI-64 based /96, this is avoided and, with a MAY in the draft, the somewhat risky choice you have described remains possible.

> That said, we have been pushing this draft for a year trying to drive concensus. If we do not limit the scope to the core objective we will be here for another year while the issue of ipv4-only hosts and apps delays and hinders broader ipv6 deployment.
> 
On this, we agree.

Still convinced to be one of the best supporters of a good 464XLAT without undue delay,
Best regards,
RD

> CB
> > _______________________________________________
> > v6ops mailing list
> > v6ops@ietf.org
> > https://www.ietf.org/mailman/listinfo/v6ops
> >
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops

Re: [v6ops] 464XLAT - support of /64 delegated prefixes