Re: [Softwires] Proposed Unified Address Mapping for encapsulation and double-translation

[Maoke <[e-mail]>]

hi Gang,

thanks for the reply.

2011/10/9 GangChen <[e-mail]>

> >> Yes, in padding case.
> >>
> >>
> > what do you mean with the "yes"? do you mean having a short-enough
> Rule-IPv6
> > is a prerequisite for deploying 4rd-pd?
> >
> > for example, a small ISP has its Rule IPv4 prefix such as 10/8, while its
> > total IPv6 block is a /48 or a /56 prefix only. the ISP is deploying
> > stateless transition, possibly 4rd-pd. is such an ISP surely able to pad
> > necessary part of IPv4 address in the first 64 bits?
>
> I guess I have already put some comments on that. Please see below.
>
>
do you mean you *require" every ISP surely having IPv6 prefix shorter than,
e.g., /32? may i beg you answer this question without equivocation. :)
thanks!


> > i don't think all current practices are "trying to encode the information
> in
> > the first 64bits". to my humble understanding, at least divi-pd encodes
> the
> > IPv4 full address information, which identifies a specific user, not in
> the
> > first 64 bits.
>
> Thanks for the clarification. That is why you propose to insert full
> IPv4 address?
>
>
well, i didn't propose anything new. some cases need full IPv4 addresses
while these cases are included in our discussion on the unified address
mapping.


> > in your proposal draft-chen-softwire-4v6-pd-00, you have
> >
> >   Delegated IPv6 prefix = IPv6 prefix + Add set + Port set
> >
> > where your Add-set is (32 - IPv4_pref_len) long, right? if the
> IPv4_pref_len
> > = 8 while the IPv6_prefix_len = 48, can your first 64bits identify a
> user?
> >
>

on the other hand, may you also answer this question?

thanks and regards,
maoke