Re: [Idr] Fwd: I-D ACTION:draft-pmohapat-idr-acceptown-community-01.txt

["John G. Scudder" <[e-mail]>]

On Apr 30, 2008, at 5:19 PM, Danny McPherson wrote:
>> Do you really think that a client which doesn't interoperate with  
>> 2796
>> RRs is viable in a network at all?  Especially one which also  
>> violates
>> the base spec by not checking the NEXT_HOP?
>
> I've seen it before in a real network.

And did you think that client was behaving appropriately then?  If so,  
why did you think it was OK for the client to violate the base spec?   
Also, you mentioned that your "real network" example was from ~10  
years ago, so while it's interesting... do you think such clients are  
still likely to exist *today*?

> Well, upon thinking about this a bit more, it also seems silly to
> me that a network would be architected in such a way as to allow
> problems with an 'upstream' RR to affect connectivity between
> two VRFs on the same PE.  I'd think to an operator trying to ensure
> availability would be worth the configuration overhead, even if
> some folks consider the extra 'state' a wash.

I guess different operators have different views about where in the  
optimization space they want to live.  For that matter, if the  
(redundant pair of) RRs goes south, the PE's connectivity to the rest  
of the world is broken anyway, local connectivity is just a special  
case.  I can understand why the survival of that special case when the  
rest of the world is broken may not be seen as a deal-breaker.

--John
_______________________________________________
Idr mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/idr