Stig Venaas wrote:
> I've been discussing destination address selection with some
> people, in particular how getaddrinfo() ideally should sort
> IPv4 addresses in the case that a host is placed in an
> IPv4-only network, and how this works with the algorithm in
> the draft.
Consider that an IPv4 network is nothing more than an NBMA link to an IPv6 node that
implements 6to4 and/or ISATAP and your perception of address sorting for nodes in an
IPv4-only net may change.
> My first thought was that this rule for an IPv6 host in an
> IPv4-only network with no known IPv6 routers would be enough
> to have IPv4 addresses sorted before IPv6 addresses. There
> was however some confusion in our discussion because some
> implementations have in this case an on-link default route,
> which means you actually have a route for all IPv6 addresses.
Your use of the terms on-link & default route together could be part of the confusion.
For one, the concept of a default route applies to the next hop used when there is no
obvious path for this destination, and for all on-link nodes there is an obvious path.
For the other, the next hop for the default route is by definition on-link. I can't
say without a picture what that last line means, but if an implementation believes
that all addresses exist on its link, that implementation would appear to be broken.
In an IPv4 world it might have made sense to try an arp for an off-link prefix since
most nodes only have one useful address, but in an IPv6 world it makes little sense to
try ND for an off-link prefix when the link-local address should work and is the
appropriate scope.
> Or does anyone think that we actually should use this
> algorithm then as well; preferring RFC 1918 destination
> for 1918 source etc?
The intent in the draft is that 1918's get treated as site local, while
draft-ietf-zeroconf-ipv4-linklocal-03.txt are treated as link-local. This sets up the
IPv4 space with the same scopes so the addr-select rules apply equally. If the IPv6
scopes don't match but the IPv4 ones do then IPv4 should be used.
Tony
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
