Re: [v6ops] SLAAC renum: Problem Statement & Operational workarounds

[Philip Homburg <[e-mail]>]

>    Actually I do not believe this is correct behavior.   Let us
>    assume prefix delegation.   If we have prefix delegation, then
>    when the CPE comes back from a power cycle, it should reconfirm
>    the prefix it had previously; the assumption is that that prefix
>    is still valid.  

This is not required behavior for a DHCP client. Traditionally, a client
does not have to write a lease it received in persistent storage. In 
fact, in the case of DHCPv6-PD in some cases that actually triggers bugs
on the ISP side and delays obtaining a new lease.

Now I'm not opposed to trying to renew the previous lease. It makes the
whole thing more complex and probably more fragile, but it is the right
thing to do.

However, the thing that actuualy makes it work (and no need for SEND there)
is to have the CPE to advertise a stale prefix with lifetimes zero to
inform hosts that this prefix no long works.