RE: : application ID and no end ...

Hi,

thanks to the many comments.
I realize that the argument of realm based routing is not applicable to ASR.

This does not automatically answer my original question which Application ID is applicable - 0 or the ID of the new application.

I suspect that this is a grey zone in the standards, and perhaps something that could by clarified in a possible RFC 3588bis.
Right now, clients should probably accept both alternatives in incoming messages.

However, I heared some arguments that ID 0 might be the better choice if messages are constructed, and a new application could clarify that this value is applicabe.

Thomas

> -----Original Message-----
> From: Loughney John (Nokia-NRC/Helsinki) 
> [mailto:[email protected]] 
> Sent: Tuesday, November 29, 2005 7:07 PM
> To: [email protected]; Belling, Thomas
> Cc: [email protected]; [email protected]
> Subject: RE: [AAA-WG]: application ID and no end ...
> 
> 
> I agree.
> 
> John
> 
> -- This email was sent from my Nokia phone --
> 
> > -----Original Message-----
> > From: [email protected] on behalf of ext Anders Kristensen
> > Received: Tue Nov 29 19:24:47 EET 2005
> > To: Belling, Thomas
> > Cc: [email protected], [email protected]
> > Subject: Re: [AAA-WG]: application ID and no end ...
> > 
> > 
> > 
> > Belling, Thomas wrote:
> > 
> > > Hi all,
> > > 
> > > I agree that in most practical scenarios realm based 
> routing would not be used for an ASR.
> > > However, this is nowhere stated on RFC 3558 - so we 
> should not rule that out.
> > 
> > Well, 3588 lists the Destination-Host AVP is being required in ASR. 
> > Doesn't that imply "peer based", as opposed to realm based, routing?
> > 
> > Anders
> > 
> > > 
> > > Furthermore, instead of demading complicated 
> considerations in this respect from each application 
> designer, wouldn´t it be easier to state that as a rule the 
> app-ID should be used for all commands within a Diameter 
> session related to that application?
> > > 
> > > Thomas
> > > 
> > > 
> > >>-----Original Message-----
> > >>From: [email protected] [mailto:[email protected]] 
> > >>On Behalf Of [email protected]
> > >>Sent: Tuesday, November 29, 2005 1:53 PM
> > >>To: [email protected]
> > >>Subject: RE: [AAA-WG]: application ID and no end ...
> > >>
> > >>
> > >>Hi,
> > >>
> > >>
> > >>>ASR is a common diameter message applicable to any 
> > >>>application and since the Destination-Host is used for 
> > >>>routing I don't see how the message can reach the wrong 
> > >>>client. So, ASR can use application id '0'IMHO.
> > >>
> > >>I think so too.
> > >>
> > >>It seems that in some cases the application-id should
> > >>be changed even if there is no changes to the command
> > >>but in other cases this is not needed.
> > >>
> > >>I think it might be difficult to find a rule that can
> > >>be applied in all cases.
> > >>
> > >>I think the main thing is that the Diameter Application
> > >>defines what app-id is used in each case taking in account
> > >>the possible error cases etc.
> > >>
> > >>
> > >>BR,
> > >>Mikko
> > >>
> > >>
> > >>
> > >>>-----Original Message-----
> > >>>From: [email protected] [mailto:[email protected]] 
> > >>>On Behalf Of ext STURA Marco Consultant
> > >>>Sent: 29 November, 2005 14:21
> > >>>To: Loughney John (Nokia-NRC/Helsinki); 
> > >>>[email protected]; [email protected]
> > >>>Cc: [email protected]
> > >>>Subject: RE: [AAA-WG]: application ID and no end ...
> > >>>
> > >>>John,
> > >>>
> > >>>but ASR is mainly sent by a server to a specific client using 
> > >>>the Destination-Host. The server must know where the Diameter 
> > >>>session to be torn down is actually running, in which client.
> > >>>
> > >>>ASR is a common diameter message applicable to any 
> > >>>application and since the Destination-Host is used for 
> > >>>routing I don't see how the message can reach the wrong 
> > >>>client. So, ASR can use application id '0'IMHO. 
> > >>>
> > >>>Regards
> > >>>Marco
> > >>>
> > >>>-----Original Message-----
> > >>>From: [email protected] [mailto:[email protected]] 
> > >>>On Behalf Of [email protected]
> > >>>Sent: Tuesday, November 29, 2005 1:06 PM
> > >>>To: [email protected]; [email protected]
> > >>>Cc: [email protected]
> > >>>Subject: RE: [AAA-WG]: application ID and no end ...
> > >>>
> > >>>Thomas,
> > >>> 
> > >>>If you are trying to reach a server that implements the new 
> > >>>application, then any commands that form part of that 
> > >>>application should use the new Application ID.  If the new 
> > >>>Application ID is not used, then the messages could not be 
> > >>>routed properly.
> > >>> 
> > >>>Of course, the drawback for this is, all relays and proxies 
> > >>>would need to be updated to enable them to route messages for 
> > >>>the new Application ID.  This is why it is not recommended to 
> > >>>create new applications unless absolutely necessary.
> > >>> 
> > >>>John
> > >>
> > > 
> > 
> > 
> > 
> 
>