Re: Should we address conflicting addresses on single MT interface?

[Alex Zinin <zinin at psg.com>]

Jeff,

Sorry for the delay--catching up.

Monday, April 5, 2004, 4:00:34 PM, Jeff Parker wrote:
> Alex -
> 	quoting from 
> http://www.ietf.org/proceedings/03jul/I-D/draft-ietf-isis-wg-multi-topology-
> 06.txt

>>> 12.2.2. Multiple MTs share an interface with overlapping addresses

>     Some additional mechanism is needed to select the correct RIBs
>     for the incoming IP packets to determine the correct RIB to make
>     a forwarding decision. For example, if the topologies are
>     QoS partitioned, then the DSCP bits in the IP packet header can
>     be utilized to make the decision. Some IP header or even packet
>     data information may be checked to make the forwarding table
>     selection, such as source IP address in the header can be used
>     to determine the desired forwarding behavior.

>>>     The generic approach of packet to multiple MT RIB mapping over
>>>     the same inbound interface is outside the scope of this draft.

> I assume that the motivation to sort this out would be to provide 
> interoperable implementations.

Exactly. While not limiting other options, providing the "feature set"
that would ensure interop.

> The draft defines a mechanism, and 
> leaves open the definition of particulars in this case.  

> Is this something that needs to be decided now?  Are
> there competing interpretations?  Are there competing
> demands for this functionality?  I see the value in
> having agreement, but wonder if we have enough experience
> with the idea to know how ISPs would want to use it tomorrow.   

What I'll do is ask the ops-dir to give them a heads-up and see if there are any
strong opinions there and then bring this to the rtg area mailing list.

Alex