RE: [dhcwg] Renumbering Requirements for Stateless DHCPv6

[[e-mail]]

Changming, 

>  For unplanned change, how can you set the lifetime? The 
> solution should work for both planned and unplanned changes. 
> The unplanned changes are more demanding. Please see the 
> scenario described in my previous email.

I don't agree at all. There is no requirement to work in the event of
unplanned events - it would be nice, but I don't agree that it is
required. There will always be breakage in the case of unplanned events,
which is precisely why operators try to avoid them. I think we are
deluded if we think that there is a solution out there that could be
guaranteed to work in the case of an unplanned event. Given that,
lifetimes are necessary and sufficient in my opinion.

> For reconfig, we already have one defined, but unicast only. 
> So the infrastruture should be there.

Why require infrastructure when it isn't necessary?
 
>  It only problem is it's unicast which does not scale well. 
> The problem is well defined in this draft and no need to repeat here.

Unicast doesn't scale and multicast is difficult to secure - that's the
bind that reconfig is in.

Mat

_______________________________________________
dhcwg mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/dhcwg