Re: [Dime] Comments on draft-ietf-dime-erp-03.txt

> Section 4 says:
>   If there is an ER server in the same domain as the authenticator
>   (local domain), Diameter routing MUST
> 
>   QUESTION:
>      Should this say "SHOULD: instead of "MUST"?
> 
>   be configured so that this ERP/DER message reaches this server, even
>   if the Destination-Realm is not the local domain.
> I wonder what the purpose of this paragraph might be: why would it be
> necessary to configure Diameter routing at all?  Actually, it seems like
> this passage could break the ERP app, since the very next paragraph says
> that the message should be routed to the bootstrapped ER server in the
> re-authentication case.  This seems like the correct behavior, so why modify
> it?

[Qin]: Not sure the ER server can terminate the ERP/DER message based on *Diameter ERP application*, 
when the Destination-Realm of the message received from the peer is not the local domain.
In my understanding, Normally Destination-Realm or Realm part of NAI is chosen by the peer and determine 
how the Diameter message is routed. Maybe the destination-Realm can be changed by the Diameter intermediate node.
But the only example I can image is Dedicated NAI is used, the Diameter proxy in the visited realm can replace the
destination-realm of message with home realm name rather than vistited realm name or local realm name.
What am I missing?

_______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime