>>>>> On Sun, 5 Sep 2004 19:17:47 -0400,
>>>>> "Manfredi, Albert E" <[EMAIL PROTECTED]> said:
>> (I would personally not use this type of "semi-auto" manual
>> configuration in this case though).
> Don't know about semi-auto. "Manual configuration" is a perfectly
> clear way of explaining that you can't guarantee address uniqueness
> in the local subnet if you can't use all of the bits of the
> supposedly globally unique number (MAC address or what have
> you). But there are ways other than manual configuration to select a
> number (maybe even at random), and then test it in the local subnet
> for uniqueness. Unless, for some reason, in this context, that won't
> work. I was wondering if I was missing something.
I'm not really sure if I understand your point, but what RFC2462 and
rfc2462bis are trying to say is, in my understanding, that if
interface_id_len + link_local_prefix_len > 128 then we cannot use the
procedure described in section 5.3 of RFC2462 for creating a
link-local address. Any other attempts to create a different
link-local address (using a random ID or a shorten IFID, or whatever)
are regarded as "manual configuration" in this context, and it is
beyond the scope of RFC2462/2462bis to describe the details of
alternative "manual" approaches.
At least to me this is pretty clear, and, specifically, I don't see
any need for changing rfc2462bis further on this matter. If you want
to introduce a change in rfc2462bis for this point, please explicitly
say so (and concrete suggested text would be highly appreciated in
this case).
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
