On Wed, 2005-03-16 at 23:58 +0200, Pekka Savola wrote: >On Wed, 16 Mar 2005, Manfredi, Albert E wrote: >> Are RFC 2893 Para. 5.2 and 5.3 going to be updated accordingly? >> Otherwise, I have no objection. > >RFC2893 is going to be obsoleted any day now, by >draft-ietf-mech-v2-xx, so this is not an issue. > >While I would have liked to remove the mention of compatible addresses >completely, deprecation as proposed by Bob is good enough for me. > >Some might argue (and argued in the past) that there should be some >health warnings about the use of mapped addresses (e.g., a reference >to now-published RFC4038), but IMHO that kind of text may be ill fit >to the address architecture and is opening a can-of-worms that we >_don't_ want to touch here. But I could live with adding a reference >if required.
I am all in for the current proposal of deprecation. The only addition I would _suggest_ is that both the mapped and compatible addresses are basically a way of representing addresses to a user. Users are familiar with IPv4 addresses (eg 192.0.2.0), thus maybe, for the people who want to use IPv6 sockaddr's to store also IPv4 addresses, they might want to represent IPv4 addresses using the normal 192.0.2.0, the on-the-wire traffic is IPv4 also in these cases* and the user sees a IPv4 address, which is what she expects. Showing ::1.2.3.4 or ::ffff:1.2.3.4 only confuses them as they would think "that is IPv6, it looks like IPv4 but it has a ':' thus it is IPv6" even though it is not. * = there where some mentioning that they used either or both compat+ mapped addresses but afaik they did not use them on the wire, eg really sticking ::ffff:1.2.3.4 in a IPv6 packet. Greets, Jeroen
signature.asc
Description: This is a digitally signed message part
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
