Re: [mif] WG Review: Multiple InterFaces (mif)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Excerpts from Margaret Wasserman on Wed, Apr 22, 2009 10:29:07AM -0400:
> Lars Eggert wrote:
>> On 2009-4-22, at 2:19, Christian Vogt wrote:
>>> It seems that folks are considering two related, yet still orthogonal
>>> topics for inclusion in the MIF charter:
>>>
>>> - Conflicts between configuration parameters.
>>>
>>> - Issues with address selection.
>>
>> I agree that both of these are important and should be worked on (and  
>> with the rest of your email, basically).
>>
>> The first one is what I thought MIF would be focusing on, as an INT WG  
>> is IMO the right venue for this.
>>
>> The second one is also important, but much more tricky, because it  
>> ties in with transports and applications (as Keith and others have  
>> pointed out already). Topics that cross area boundaries are always a  
>> bit difficult to charter. I'm at this point not fully convinced that  
>> simply throwing this in with topic #1 into one WG is going to work.
>
> I disagree with your conclusion for two reasons:
>
> (1) As I pointed out in my previous message to Christian, address  
> selection is not (today) a transport-layer or application-layer function  
> in most cases.  Given that this is currently an Internet-layer function,  
> I think it makes sense to analyze the issues with address selection (as  
> part of the whole address/interface/router selection process)  in an  
> Internet Area group.  If we find that one of the problems we have is  
> that the Internet layer doesn't have the right information to make these  
> decisions, then possibly some follow-on work might need to be chartered  
> elsewhere.
>
> (2) There is no way that these decisions can be made solely at the  
> transport or application layer, because source (and to a lesser degree  
> destination) address selection is tightly tied to the first-hop  
> forwarding decision.  The outbound interface, source address and default  
> router all have to be selected in a coordinate process, to avoid sending  
> traffic that will be discarded on the outbound path, due to router 
> filters.
>
> So, while agree that address selection affects transport layers and  
> applications, and that it might be necessary for transport layers and  
> applications to have better ways of influencing it, I do not believe  
> that address selection is a transport layer or applications layer  
> function today, nor do I think it can be done solely at those layers in  
> the future.
>
> Margaret

The first problem (configuration conflicts) is best dealt with before
the second problem (address selection) even arises.  First I
initialize my interface, and then I use it.  They impinge on each
other but they are not tightly bound to each other -- they do not have
to be worked on in the same Working Group.  

I suggest chartering MIF to focus on problem #1, let it make progress,
and in the meantime figure out how to organize work on problem #2 in
parallel.

Scott
_______________________________________________

Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]