14:01 < amd2-ripe> I feel sorry for IPv4 that doesn't have it's own working group, they're missing out!
14:01 < Anand_RIPENCC> Jen Linkova has opened the session
14:03 < Anand_RIPENCC> Geoff Huston is presenting on "Is IPv6 Only for the Rich?"
14:22 < Anand_RIPENCC> Any comments or questions for Geoff?
14:30 < Anand_RIPENCC> Helge Holz is presenting on "Painting by Numbers: an Update"
14:46 < Anand_RIPENCC> Any questions or comments for Helge?
14:54 < Anand_RIPENCC> Jordi Palet is presenting on "/64 per Host"
14:57 < Anand_RIPENCC> Any comments or questions for Jordi?
15:00 < Anand_RIPENCC> Jen Linkova is presenting on "Enterprise IPv6 Multihoming using PA"
15:18 < TimB> I've tried to do this is my org on office networks. They theory works, but clients don't work very well in practice.
15:18 < Anand_RIPENCC> Any comments or questions for Jen?
15:19 < Anand_RIPENCC> TimB: should I pass on your message to the speaker?
15:19 < TimB> lots of clients don't cope at all PIO preferred_lifetime = 0
15:19 < TimB> Anand, yes please
15:19 < TimB> Tim Bray, ProVu Communications Ltd
15:27 < TimB> Mainly linux desktops that broke.
15:27 < TimB> on wired ethernet
15:29 < TimB> For medium sized companies, with any kind of onsite fileservers or stuff, the PI space is much much better
15:30 < cougar> I see that the problem is that all local nodes stop using one prefix and need to switch to another and this causes a break even if there is no problem in local network
15:31 < cougar> is it something can could be solved wit h using ULA internally?
15:31 < TimB> cougar, when we tried this. We kept finding things like we couldn't print when prefixes changed.
15:32 < TimB> I think Jen's approach is fine for really simple small networks, if all the clients can be tested to work really nicely. If anythink breaks, might as well got back to NATTED IPv4