Re: [Softwires] MAP Open issues: Interface-id

hi,

On 11/8/2011 5:08 PM, GangChen wrote:
> ...
>> the consequence of that is that the destination IPv6 address will change for
>> every flow.
>> the MAP node cannot any longer listen to a single IPv6 address for MAP
>> traffic, but has to intercept packets for a whole prefix.
> Yes. I see there are two choices when MAP nodes process incoming packages
> i) Should we let MAP node interface bind to a single static IPv6 address
Well, by an interface-id for MAP, works.
> ii) Should we let MAP node interface bind to a route to
> IPv4-translatable IPv6 prefix
This is what I ever proposed, a new interpretation of the V/U octet, 
forming a specific route, say /72 pointing to the MAP function.

>
> In translation case, above two cases are possible. That is
> implementation specific. So I suggest to put above design as int-id
> candidate for more inputs from the community
But the variable destination addresses to target MAP function makes the 
above impossible. We can't ignore this issue during the MAP design.


Cheers,
Jacni
>
> BRs
>
> Gang
> _______________________________________________
> Softwires mailing list
> Softwires@ietf.org
> https://www.ietf.org/mailman/listinfo/softwires
>