On Sat, Mar 17, 2012 at 2:22 AM, Dave Thaler <[email protected]> wrote: > Brian Carpenter writes: > [...] >> Let me be clear. If a local service has (for some reason) both a ULA and a >> non- >> ULA global address, and the host has both, I think the correct default >> behaviour is for the ULA address pair to be used. > > As I put into the doc, I don't think that's quite right. > > If both the source and dest ULAs are in the same /48 then I think the correct > default is as you say (use ULA). > > If the source and dest ULAs are in different /48's then I think the correct > default is instead to use the non-ULA global, since there's no guarantee of > routability between different /48s. So unless configured otherwise, one > has to assume it's far more problematic than a non-ULA global. > Do you mean "no guarantee of symmetric routability"? The fact that the packet arrived in the first place seems to indicate earlier policy choices (e.g. the sender may not have a non-ULA global address, and the two /48s already seem to share a common definition of "site").
I am still relatively new to homenet and I am surely missing a lot of background. Has anyone discussed dealing with multiple /48 ULA prefixes in a single site? Thanks, -K- > You'll find the above logic in the current 3484bis draft. > > -Dave > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > [email protected] > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
