Re: [Dime] Diameter Group: Type?

I agree with Avi, I think having the "grouped" information as a type
makes more sense than as a flag.
If the AVP code/vendor combination is unknown in the local dictionary,
there is no point in being able to "parse" the inside AVPs -- and I
think doing so (which would be possible with a "G"rouped flag in AVP
header) would actually do more harm than good.
As you wrote, parser's performance is a resource not to be wasted.
Therefore, what is the point of parsing the "inside" level of AVPs when
you don't understand the semantics of the grouped AVP that surrond them ?

Best regards,
Sebastien.

Le 25/01/2011 07:48, Avi Lior a écrit :
> Well, Lets say you are right.  Don't you think that this is a bit late?
>
> But I don't agree with you that a Group is not a kind of type and
> hence it could be a basic type (if you get my meaning)
>
> Could you explain why do you think that a flag will simplify parsing?
>
> -- Avi Lior
> --Bridgewater Systems
>
>
> From: Erez Nassimi <erez.nassimi@amdocs.com
> <mailto:erez.nassimi@amdocs.com>>
> Date: Mon, 24 Jan 2011 08:51:27 -0500
> To: "dime@ietf.org <mailto:dime@ietf.org>" <dime@ietf.org
> <mailto:dime@ietf.org>>
> Subject: [Dime] Diameter Group: Type?
>
> To whom it may concern:
>
> ======================
>
>  
>
> Dealing with the Diameter standard for more than 2 years, I always had
> the following question, I never dared to ask. But even after 2 years,
> I do not seem to have a good answer for it:
>
>  
>
> In RFC 3588, section 4.2 – AVP Data Formats, “Grouped” is defined as a
> “basic type”. Doesn’t it make more sense to set a flag for Grouped AVP
> in flags section? Group may be basic but it is not areal type. It’s
> just an indication of a complex data structure.
>
>  
>
> If “Grouped” is defined as a flag, it will simplify any parser’s work
> – and performance is such a rare resource.
>
>  
>
> Thanks in advance,
>
> *Erez Nassimi*
>
> *erezna@amdocs.com <mailto:erezn@hotmail.com>***
>
> *Desk : +972-9-77-86073*
>
> *Cell : +972-54-7296230***
>
> *Fax  : +972-9-77-61783*
>
>  
>
> */Did you know…?/*
>
> With Amdocs’ innovative new “Turbo Charging”
> <http://www.amdocs.com/Site/News/News+Articles/2008/Press+Releases/Turbo+Charging.htm>
> complex event processing technology, service providers can process
> thousands of charging events per CPU in real-time over low-cost
> hardware (like blade servers), all with the comprehensive
> functionality of Amdocs CES – Billing 7.5.
>
>  
>
> This message and the information contained herein is proprietary and confidential and subject to the Amdocs policy statement,
> you may review at http://www.amdocs.com/email_disclaimer.asp
>
>
> _______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime

-- 
Sebastien Decugis
Research fellow
Network Architecture Group
NICT (nict.go.jp)