On 04-Jul-19 09:03, Pascal Thubert (pthubert) wrote:
> Sorry Brian I miss your point.
> 
> The discussion here is to find a way for the node to autoconf an address and 
> then notice the router(s) proactively so the ND NCE is ready when the first 
> packet comes from the outside for this node.
> 
> I agree wholeheartedly with Jen’s requirement.

Yes, I am not against satisfying this requirement, but we have a considerable 
legacy of stacks that don't do this - so until that legacy has vanished, the 
routers need to deal with it somehow.

> I agree less with the idea of overstretching ND as it stands to achieve this, 
> for reasons I gave already. So prepopulate the NCE, yes, but do it right with 
> protocol elements that guarantee a state that is accurate, secure, and 
> persistent for a committed lifetime, not with yet another patch to the old 
> structure.
> 
> I’m interested to have a parallel discussion on where RFC 8505 can not apply. 
> In the products and use cases I’m aware of, it could, since we are actually 
> faking it by snooping ND and DHCP to achieve similar but less accurate 
> results.

So if you are advocating a generalisation of RFC8505 to non-6lo LANs, that's 
certainly a discussion we could have, IMHO.

Regards
    Brian

> 
> Take care,
> 
> Pascal
> 
>> Le 3 juil. 2019 à 22:39, Brian E Carpenter <[email protected]> a 
>> écrit :
>>
>> On 03-Jul-19 20:13, Pascal Thubert (pthubert) wrote:
>>
>> ...
>>> I'm baffled that the reactive ND is still the official technique for IPv6 
>>> lookup at 6MAN.
>>
>> How can it be otherwise when a node can give itself a new address at any 
>> time without notice?
>>
>> I'm not arguing with you about RFC 6775/8505 networks, but that doesn't 
>> apply everywhere.
>>
>>    Brian

_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch

Reply via email to