Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

"TROTTIN, JEAN-JACQUES (JEAN-JACQUES)" <jean-jacques.trottin@alcatel-lucent.com> Thu, 27 March 2014 14:35 UTC

Return-Path: <jean-jacques.trottin@alcatel-lucent.com>
X-Original-To: dime@ietfa.amsl.com
Delivered-To: dime@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 827DF1A068A for <dime@ietfa.amsl.com>; Thu, 27 Mar 2014 07:35:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.899
X-Spam-Level:
X-Spam-Status: No, score=-6.899 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id djm2LuD0lXcu for <dime@ietfa.amsl.com>; Thu, 27 Mar 2014 07:35:37 -0700 (PDT)
Received: from hoemail1.alcatel.com (hoemail1.alcatel.com [192.160.6.148]) by ietfa.amsl.com (Postfix) with ESMTP id CF0441A06C1 for <dime@ietf.org>; Thu, 27 Mar 2014 07:35:36 -0700 (PDT)
Received: from fr711usmtp1.zeu.alcatel-lucent.com (h135-239-2-122.lucent.com [135.239.2.122]) by hoemail1.alcatel.com (8.13.8/IER-o) with ESMTP id s2REZWog029900 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL) for <dime@ietf.org>; Thu, 27 Mar 2014 09:35:34 -0500 (CDT)
Received: from FR712WXCHHUB03.zeu.alcatel-lucent.com (fr712wxchhub03.zeu.alcatel-lucent.com [135.239.2.74]) by fr711usmtp1.zeu.alcatel-lucent.com (GMO) with ESMTP id s2REZRmS022198 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL) for <dime@ietf.org>; Thu, 27 Mar 2014 15:35:28 +0100
Received: from FR712WXCHMBA12.zeu.alcatel-lucent.com ([169.254.8.152]) by FR712WXCHHUB03.zeu.alcatel-lucent.com ([135.239.2.74]) with mapi id 14.02.0247.003; Thu, 27 Mar 2014 15:35:25 +0100
From: "TROTTIN, JEAN-JACQUES (JEAN-JACQUES)" <jean-jacques.trottin@alcatel-lucent.com>
To: "dime@ietf.org" <dime@ietf.org>
Thread-Topic: [Dime] [dime] #54: OC-Report-Type as mandatory AVP
Thread-Index: AQHPQsn+hW8mxnR4ZUW/UfMU7zBoKJrvnpVggAAIHgCAAbwsIP//2cWAgAGGpGD//6iMAIAAiQOg//99GVAAAQlQAAARdy8Q///D9QD//3OVcIAApYqA//5hYeD//Jsx8ADZeeeA///qNtD//8n+gP//j8UA//7p8eD//cwccA==
Date: Thu, 27 Mar 2014 14:35:24 +0000
Message-ID: <E194C2E18676714DACA9C3A2516265D20267B3A9@FR712WXCHMBA12.zeu.alcatel-lucent.com>
References: <075.72da31b401c033905a4fb81d09a8b4aa@trac.tools.ietf.org> <53287893.5020203@usdonovans.com> <087A34937E64E74E848732CFF8354B920979DDA2@ESESSMB101.ericsson.se> <53288284.5040606@usdonovans.com> <A7A9CDB7-9D90-458B-8C24-F0BFF52F897C@gmail.com> <26C84DFD55BC3040A45BF70926E55F2587C3D80D@SZXEMA512-MBX.china.huawei.com> <53302446.9080700@usdonovans.com> <26C84DFD55BC3040A45BF70926E55F2587C3DE93@SZXEMA512-MBX.china.huawei.com> <533178CD.9030707@usdonovans.com>, <26C84DFD55BC3040A45BF70926E55F2587C3E4C3@SZXEMA512-MBX.china.huawei.com> <11673_1395816227_53327723_11673_1919_1_4wjyrmeak04xst2onmes7ybh.1395816218573@email.android.com> <26C84DFD55BC3040A45BF70926E55F2587C3E57B@SZXEMA512-MBX.china.huawei.com> <26C84DFD55BC3040A45BF70926E55F2587C3E5AD@SZXEMA512-MBX.china.huawei.com> <1369_1395819319_53328337_1369_10404_1_6B7134B31289DC4FAF731D844122B36E51C25A@PEXCVZYM13.corporate.adroot.infra.ftgroup> <26C84DFD55BC3040A45BF70926E55F2587C3E671@SZXEMA512-MBX.china.huawei.com> <5332C60E.2060805@usdonovans.com> <26C84DFD55BC3040A45BF70926E55F2587C3E932@SZXEMA512-MBX.china.huawei.com> <14209_1395841826_5332DB22_14209_9620_1_6B7134B31289DC4FAF731D844122B36E51E2FD@PEXCVZYM13.corporate.adroot.infra.ftgroup> <26C84DFD55BC3040A45BF70926E55F2587C3ED23@SZXEMA512-MBX.china.huawei.com> <22839_1395912677_5333EFE4_22839_3586_1_6B7134B31289DC4FAF731D844122B36E520AC9@PEXCVZYM13.corporate.adroot.infra.ftgroup> <26C84DFD55BC3040A45BF70926E55F2587C3EEEF@SZXEMA512-MBX.china.huawei.com> <11312_1395916348_5333FE3C_11312_7917_5_6B7134B31289DC4FAF731D844122B36E520C95@PEXCVZYM13.corporate.adroot.infra.ftgroup> <26C84DFD55BC3040A45BF70926E55F2587C3EF92@SZXEMA512-MBX.china.huawei.com> <5BCBA1FC2B7F0B4C9D935572D9000668151D8B92@DEMUMBX014.nsn-intra.net>
Accept-Language: fr-FR, en-US
Content-Language: fr-FR
X-MS-Has-Attach:
X-MS-TNEF-Correlator:
x-originating-ip: [135.239.27.41]
Content-Type: multipart/alternative; boundary="_000_E194C2E18676714DACA9C3A2516265D20267B3A9FR712WXCHMBA12z_"
MIME-Version: 1.0
Archived-At: http://mailarchive.ietf.org/arch/msg/dime/MUO7rrf-93EBYrn3JcHg-O4gHEY
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP
X-BeenThere: dime@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Diameter Maintanence and Extentions Working Group <dime.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/dime>, <mailto:dime-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/dime/>
List-Post: <mailto:dime@ietf.org>
List-Help: <mailto:dime-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/dime>, <mailto:dime-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 27 Mar 2014 14:35:47 -0000

Hi

I refer to my yesterday previous mail.
People have expressed

-          A preference. On my side,  I have a preference for the existing text with  non mandatory  AVP. But I can accept the mandatory AVP as I already said

-          Or a strong concern in favor of one solution, this is the case of Susan. I asked if there were voice who have a strong concern  in favor of the mandatory AVP .  There were was no answer up to now  and I am not aware of voice with a strong concern  in favor of the mandatory .



Best regards

JJacques

De : DiME [mailto:dime-bounces@ietf.org] De la part de Wiehe, Ulrich (NSN - DE/Munich)
Envoyé : jeudi 27 mars 2014 12:53
À : ext Shishufeng (Susan); lionel.morand@orange.com<mailto:lionel.morand@orange.com>; Steve Donovan; Jouni Korhonen
Cc : dime@ietf.org<mailto:dime@ietf.org>
Objet : Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

[LM] let see if there are other voice that could not accept the proposed approach.
Susan: Ok, but to be fair, maybe we can also ask who could not accept this AVP as optional. As I know, Steve could live with it as optional.

I can accept both (though I have a slight preference in favour of Susan).
Actually I think this is a so minor minor minor issue that it  could serve as a very good opportunity for people to show their willingness to give in, in order to make progress.

Ulrich

From: DiME [mailto:dime-bounces@ietf.org] On Behalf Of ext Shishufeng (Susan)
Sent: Thursday, March 27, 2014 12:38 PM
To: lionel.morand@orange.com<mailto:lionel.morand@orange.com>; Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hello Lionel,

Please see below.

Best Regards,
Susan

From: lionel.morand@orange.com<mailto:lionel.morand@orange.com> [mailto:lionel.morand@orange.com]
Sent: Thursday, March 27, 2014 6:32 PM
To: Shishufeng (Susan); Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Subject: RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hi Susan,

Please see below.

Lionel

De : Shishufeng (Susan) [mailto:susan.shishufeng@huawei.com]
Envoyé : jeudi 27 mars 2014 11:04
À : MORAND Lionel IMT/OLN; Steve Donovan; Jouni Korhonen
Cc : dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Objet : RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hello Lionel,


-          I'm not talking the specific optional AVPs defined in this draft, but others e.g. defined in 3GPP specifications. We have a lot optional AVPs with default values.
[LM] in 3GPP there are introduced in existing applications and usually in commands. There is no other options to include them are purely optional.
Susan: Not exactly, we have lots of optional AVPs from the first release.


-           my understanding, some cases an AVP is needed, some cases it is not needed, that's the key point to justify its optionality.
[LM] I agree that we don't have the same understanding here.


-          Apart from the benefit of less AVPs and less error cases with this AVP as optional, another key point is that the server(host) will never care about the AVP at all. It will never need to remember to put such an AVP into the message. Only agent needs to add such an AVP in case putting into realm report, this would make things much simpler from my point of view.
[LM] that is wrong to say "never". A server can have realm-based information that can be provided to reacting nodes. It is only depending of the server implementation and environment. For instance, if a realm is entirely dedicated to a type of server, a farm based implementation would provide such functionality, as discussed.
Susan: In this case the server behaves more like an agent, not a pure server, i.e. the server provides agent functionality. In IETF, server and agent has clear difference.

Further, the discussion on Realm is still ongoing and whether it can work well is still under investigation.
[LM] I think that the ongoing discussion is too tight to specific use cases, important ones for 3GPP I admit. However the whole goal of this work is to have something that it will work in any case. If there are some requirements for realm-based info, I don't see a specific reason to restrict the use of report type. And don't forget that new report types can be created in the future. So it is not restricted to Host vs realm.
Susan: May I ask what's wrong with this not always needed AVP as optional?

The benefit to have it as optional is clear, and this would not bring any harm. I hope this can be reconsidered.
[LM] let see if there are other voice that could not accept the proposed approach.
Susan: Ok, but to be fair, maybe we can also ask who could not accept this AVP as optional. As I know, Steve could live with it as optional.

Best Regards,
Susan

From: lionel.morand@orange.com<mailto:lionel.morand@orange.com> [mailto:lionel.morand@orange.com]
Sent: Thursday, March 27, 2014 5:31 PM
To: Shishufeng (Susan); Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Subject: RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hi Susan,

In this specific case, the info carried by the AVP is not optional. You need this info in the algorithm as it will modify the behavior of the reacting node.
The current solution is to define a default value when not received.

I believe that at the protocol level, having this AVP in every request is cleaner than relying on default value when not present. The initial goal for such default value was to limit the number of AVPs transmitted over the interfaces. And I think that it is not so relevant according to the output of the other discussions and such optimization is not deemed required anymore.

Moreover, the default value today would be likely "Host" but this is only based on current working assumptions for some 3GPP interfaces. When used in some networks, Realm is perfectly valid as default value if the system is designed for that. I think that AT&T and Verizon have clearly expressed their requirements for that.

This reasoning leads to change the existing for Grouped AVP from:



OC-OLR ::= < AVP Header: TBD2 >

           < OC-Sequence-Number >

           [ OC-Report-Type ]

           [ OC-Reduction-Percentage ]

           [ OC-Validity-Duration ]

         * [ AVP ]
To:


OC-OLR ::= < AVP Header: TBD2 >

           < OC-Sequence-Number >

           { OC-Report-Type }

           [ OC-Reduction-Percentage ]

           [ OC-Validity-Duration ]

         * [ AVP ]

I would like to clarify that this discussion is totally independent of the M-bit setting of the AVP.

And about why not put the other AVPs as required in the Grouped AVP, it is because it is abvious that the other ones are linked to the reduction algo defined in this document. For other algos, you could find other AVPs (relying on the extensibility of this Grouped AVP) instead.

Let me know if the clarification is useful.


Regards,

Lionel

De : Shishufeng (Susan) [mailto:susan.shishufeng@huawei.com]
Envoyé : jeudi 27 mars 2014 07:38
À : MORAND Lionel IMT/OLN; Steve Donovan; Jouni Korhonen
Cc : dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Objet : RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hi Lionel,

If you cares, please clarify how this decision was made based on the newly ongoing technical discussion. I really don't understand a useless AVP in most cases has to be mandatory.

May I ask why not define every optional AVP as mandatory if the only point you made decision was that it will not block anything?

Best Regards,
Susan

From: lionel.morand@orange.com<mailto:lionel.morand@orange.com> [mailto:lionel.morand@orange.com]
Sent: Wednesday, March 26, 2014 9:50 PM
To: Shishufeng (Susan); Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Subject: RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hi,

Because it was not possible to make everyone happy, I made a decision. I think that this decision is valid and will not block anything.
To be fair, I think that the majority could accept both while few have a strong preference for optional or mandatory.  And I picked one.
Can we please discuss about something else or is it the most critical point to solve? I think that everyone is waiting for a version 02 before the end of this week.

Regards,

Lionel


De : Shishufeng (Susan) [mailto:susan.shishufeng@huawei.com]
Envoyé : mercredi 26 mars 2014 13:53
À : Steve Donovan; MORAND Lionel IMT/OLN; Jouni Korhonen
Cc : dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Objet : RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hello Steve,

Ok, let's come back with technical discussion then.


-          First a not always needed AVP cannot justify why it shall be mandatory.

-          Secondly, in my view, having a not always needed AVP as mandatory cannot be less error prone. On the contrary, it would introduce more error cases. RFC 6733 defines a couple of error codes e.g. DIAMETER_MISSING_AVP 5005, DIAMETER_INVALID_AVP_VALUE 5004 and maybe others to deal with the possible error cases. All the handling with this AVP consumes the resource of both reporting nodes and reacting nodes. I don't see any benefit to have it.


Best Regards,
Susan


From: Steve Donovan [mailto:srdonovan@usdonovans.com]
Sent: Wednesday, March 26, 2014 8:21 PM
To: Shishufeng (Susan); lionel.morand@orange.com<mailto:lionel.morand@orange.com>; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

I disagree that the majority of individuals (companies don't apply here) have a preference.  We have not established the view of a majority.

If we took a vote of individuals on this, I would vote for making it explicit.

I also support Lionel being able to make a decision.  We all need to be flexible enough to accept that decision, especially when there is no technical reason to disagree with the decision.

Steve
On 3/26/14 3:03 AM, Shishufeng (Susan) wrote:
Hello Lionel,

Thanks for your reply.

It is clear that majority companies involved in the discussion prefer to not have this AVP as mandatory. And Steve is ok with it as it is, and thus everyone is ok with it as optional. I'm wondering why you choose another way forward, and don't take opinion from majority companies into account, which is strange for me.

We talked much about it, and in some major cases, this AVP is useless. If needed, anyway an optional AVP has to be present, I don't see any problem with it. We did a lot like this in 3GPP spec. I don't understand how in this case it shall be mandatory if you also agree that it is not needed in some cases.

Best Regards,
Susan


From: lionel.morand@orange.com<mailto:lionel.morand@orange.com> [mailto:lionel.morand@orange.com]
Sent: Wednesday, March 26, 2014 3:35 PM
To: Shishufeng (Susan); Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>; Benoit Claise
Subject: RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hi Susan,

I do care about comments from other but we need to move on.

My decision is based on the following: from a technical point of view, there is no issue if we define this AVP as required. From a protocol aspect, it is cleaner as a report ALWAYS applies to a given type. Default value was considered as sub-optimal by Jouni, Ben, Steve, Maria Cruz (who has triggered this issue) and myself.

Now, about the process, a 100% is not required to move forward. As it is not possible to get consensus on this issue, I use my prerogative to pick one solution, the one that seems to be the best solution at this stage, to be able to close the discussion at this stage and produce the draft that we need.

After, it is always possible to challenge again the content of the draft if there is still concern.

Benoit, the Area Director, is copied. It is not a putsch. Just administrative way agreed within IETF to be able to move forward a WG draft, which is the ultimate goal of everyone I think.

Regards,

Lionel.

De : Shishufeng (Susan) [mailto:susan.shishufeng@huawei.com]
Envoyé : mercredi 26 mars 2014 08:06
À : Shishufeng (Susan); MORAND Lionel IMT/OLN; Steve Donovan; Jouni Korhonen
Cc : dime@ietf.org<mailto:dime@ietf.org>
Objet : RE: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hello Lionel,

Further clarification:

I don't agree with this, not ok to everyone as you said.

Best Regards,
Susan

From: Shishufeng (Susan) [mailto:susan.shishufeng@huawei.com]
Sent: Wednesday, March 26, 2014 2:56 PM
To: lionel.morand@orange.com<mailto:lionel.morand@orange.com>; Steve Donovan; Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Hello Lionel,

I don't understand it.

Please clarify if you care about other people's comments.

Best Regards,
Susan

From: lionel.morand@orange.com<mailto:lionel.morand@orange.com> [mailto:lionel.morand@orange.com]
Sent: Wednesday, March 26, 2014 2:44 PM
To: Steve Donovan; Jouni Korhonen; Shishufeng (Susan)
Cc: dime@ietf.org<mailto:dime@ietf.org>
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP


As chair and temporary doc shepherd,



Please stop this thread for now.



As mandatory/required is ok for everyone (even if useless in certain case), let use it for now.



Lionel





"Shishufeng (Susan)" <susan.shishufeng@huawei.com<mailto:susan.shishufeng@huawei.com>> a écrit :


Hello Steve,

Thanks for clarifying the IETF procedure. I'm not familiar with it, while I know the draft is mainly for 3GPP use, that's why we 3GPP delegates are deeply involved in this specific discussion. If most of 3GPP people think it is not so needed I couldn't understand why it shall be mandatory.

>From technical point of view, in the case realm based report type is not needed, nothing wrong without this AVP, and even better and cleaner.

And you ever said you have preference but ok with either way forward, i.e., make it mandatory or optional. Then let's move on with the draft as it is on this point, if you agree.

Best Regards,
Susan

From: Steve Donovan [mailto:srdonovan@usdonovans.com]
Sent: Tuesday, March 25, 2014 8:39 PM
To: Shishufeng (Susan); Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Susan,

We have not been following a process of determining consensus based on a majority of companies expressing a preference.  It is also the case that, in the IETF, companies do not contribute, individuals contribute.

In addition, if we did take a "vote" on this one, I'm not sure which side would actually have a majority.

We might need to change our process to speed things up, but right now we have been striving for true consensus where everyone agrees.  Note that this doesn't mean everyone agrees with the technical reasoning behind the decision.  There have been many cases where agreement is reached because it was more important to get something finished then to win a technical argument.

If we can't start moving a little faster then we will likely need to change to rough consensus, where the measure is that most everyone agrees.  However, in the IETF, even this is not a voting process.  If things are close to 50-50 in opinions then the correct process is to continue to discuss the technical merits of each alternative until rough consensus is reached.

Regards,

Steve
On 3/25/14, 2:00 AM, Shishufeng (Susan) wrote:
Hi Steve,

As I know, majority companies expressed preference to keep the AVP as optional and keep the texts as they are. You have preference to have it explicitly but ok with either way. That's how I assumed we reached consensus.

Best Regards,
Susan

From: Steve Donovan [mailto:srdonovan@usdonovans.com]
Sent: Monday, March 24, 2014 8:26 PM
To: Shishufeng (Susan); Jouni Korhonen
Cc: dime@ietf.org<mailto:dime@ietf.org>
Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP

Susan,

We are in the middle of the discussion and have not yet reached consensus.

I agree with Jouni on making it explicit.  Either way, we should try to make a decision quickly.

Steve
On 3/23/14 10:59 PM, Shishufeng (Susan) wrote:

Hello Jouni,



I assume we had a lot of discussion on this and reached consensus to keep it as it is in the draft.



Best Regards,

Susan





-----Original Message-----

From: Jouni Korhonen [mailto:jouni.nospam@gmail.com]

Sent: Saturday, March 22, 2014 10:38 AM

To: Steve Donovan

Cc: dime@ietf.org<mailto:dime@ietf.org>

Subject: Re: [Dime] [dime] #54: OC-Report-Type as mandatory AVP





Lets have it explicit then. Use '<' and '>' to make the position fixed.



- Jouni



On Mar 19, 2014, at 1:29 AM, Steve Donovan <srdonovan@usdonovans.com><mailto:srdonovan@usdonovans.com> wrote:



I'm ok with either direction but generally lean toward being explicit.



Do we have other opinions?



Steve

On 3/18/14 12:16 PM, Maria Cruz Bartolome wrote:

Hello,

I think the agreement tendency is the contrary: OC-Report-Type is not required, while default value is Host. i.e. it will remain as it is now in the draft.

This may be of some advantage for some applications that may only use Host, as long as they may never generate Realm reports.

If there is consensus on this, I will go with this.

Best regards

/MCruz