Re: Extending a /64

[[e-mail]]

Philip,

>> If we say the link type is "point to point" and that's the only 
>> definition, it'd be applicable to PPP(oE), GTP, the 
>> point-to-point-ethernet proposal, possibly some other deployment model 
>> where it can be guaranteed that there is only a single recipient of 
>> packets. Other tunnel types also comes to mind.
>> 
>> It isn't hard to imagine other media types being designed with this in 
>> mind going forward...
> 
> In my opinion if we do anything standards track, we have include protocol
> elements and requirements that allow a downstream router to verify that the
> prefix is still valid. I.e., in my opinion we need to move away from
> lifetimes, and link state is a signal but might be an unreliable signal.

Agree with regards to lifetimes.
Addressing != Routing.
You shouldn't use a change in reachability to renumber your network.
Imagine what happens if that link flaps once a second.

> So just like ND has NUD to verify that a neighbor is still there, we need
> something to verify that prefix is still there. Ideally we can extend that
> to SLAAC and DHCP PD and have a model where changes in an upstream link
> can propagate downstream realiably.

The reachability check verifies that the next-hop router has a forwarding entry for our route.
It does not guarantee anything else.
e.g. in a multi-prefix multi-homed network, the host is the only entity that can verify end to end reachability, and it has to do it by probing (with all combinations of SA/DA).
The host stack must treat addresses as candidates, where all have potential different reachability.
Therefore, having stale addresses doesn't so much matter.

> Real point to point protocols (such as ppp) are in some sense 'easy' in that
> you know there is only one other party at the other end. Point-to-point
> ethernet is a lot more complex. It would be nice to support point-to-point
> ethernet but it might require quite a bit of work to get all corner cases
> right.


Best regards,
Ole