Re: [DNA] Some questions about Simple DNA

[Suresh Krishnan <suresh.krishnan@ericsson.com>]

Hi Bernard,

Bernard Aboba wrote:
>> By definition, for a host to do DNAv6, it has already completed
>> duplicate address detection on the address.  Why should the host
>> be required to do (Optimistic) DAD every time it awakes from sleep?
>> Given the duty cycles we are now seeing, this means that the host
>> will never obtain a usable address!
> 
> This is to deal with another host picking up the same address after the 
> host went to sleep/left the link. The new host thinks that its address 
> is unique while in fact it is not. I do not see any other way around this.
> 
> [BA] In the case of an address obtained from DHCPv6, where the lease is
> still valid, the host had previously done DAD and a compliant DHCPv6 
> server will not give out the same address to multiple hosts.  So why 
> should the host do DAD again? 

I am fine with the host not doing DAD in this case. I think that this is 
bringing out is that there is a difference based on whether the address 
was statelessly autoconfigured or acquired through DHCP.

* Address was acquired through DHCPv6 : Server does not give out the 
address to anyone else and hopefully a router is not advertising the 
same prefix as available for SLAAC. In this case it is fine to skip the DAD.

* Address was configured using SLAAC : Nobody is keeping track of the 
address used by the node. In this case when the node gets back into the 
link, it needs to do DAD.

Is this an acceptable bifurcation?

Thanks
Suresh