RE: [AVT] RTCP-XR MIB Comments (1)

["Alan Clark" <alan at telchemy.com>]

Title: RTCP-XR MIB Comments (1)

Dan

 

Thanks for the helpful comments, and I look forward to more:-)  I'll be at the IETF meeting next week, plan to try and progress the draft and would welcome some direct input from you.  I'll look through your initial feedback and try and incorporate as much as possible prior to next week.

 

With regard to the scope - this reflects the guidance given at the AVT mtg when the last draft was presented - which was that it would be more expedient to refocus from the original RTPMIBV2 to an RTCP XR MIB.  The name of the current draft was kept as RTPMIBV2-01 based on advice from the AVT chairs - I'm sure that when the draft is discussed next week then we will agree to rename it.

 

The intended applications include use in a Media Gateway, where it may be desirable to access multiple sets of call quality data through SNMP.  Personally I don't think SNMP is an ideal way to access data from individual IP phones and similar clients, other than on an occasional basis.

 

Regards

 

Alan

 

 

 

-----Original Message-----
From: avt-bounces at ietf.org [mailto:avt-bounces at ietf.org]On Behalf Of Romascanu, Dan (Dan)
Sent: Tuesday, July 27, 2004 1:50 PM
To: avt at ietf.org
Cc: bwijnen at lucent.com
Subject: [AVT] RTCP-XR MIB Comments (1)

Following are my comments on a first review of draft-clark-avt-rtpmibv2-01.txt. I am breaking it in smaller pieces, because the initial mail was bounced back by the AVT WG mail list for being too large.

I did not perform yet a full 'MIB Doctor' review of this document. There seem to be enough serious problems to fix before getting to the details with this document. In general, the guidelines for authors and reviewers of MIB documents http://www.ietf.org/internet-drafts/draft-ietf-ops-mib-review-guidelines-03.txt seem to have been ignored and I would recommend to the authors to have a serious look and implement these guidelines before the next version.

1. Title and scope - This MIB module is not a 'Proposed Real-Time Transport Protocol Management Information Base Version 2' as the title claims, but an RTCP-XR MIB. I actually believe and expressed my opinion in the past that the AVT WG should first deal with doing an updated RTP MIB, as the charter says as well, before building extensions to a MIB module that models an older version of RTP. However, the title of this I-d should reflect the content.

2. What is the model for SNMP management that this MIB module provides? If such a MIB runs in each endpoint, how would the solution scale for an Internet deployment? If the authors believe that they have an answer, they should include some examples of deployment

3. The boilerplate is not updated, and does not refer to the latest version of the SNMP standards. See http://www.ops.ietf.org/mib-boilerplate.html for the updated boilerplate

4. I think that the concept of an 'RTP monitor' needs to be explained better. It is mentioned in the Overview, but not detailed 2.1. I feel it may be important in the conception of the authors, but the definition is missing, and 2.2.2 misses a clear explanation where monitors are deployed relative to the RTP and RTCP traffic on given sessions, etc.

5. The following text in 2.3 is not clear to me:

'   rtpXrVoipIndex is a global object that permits a network-

   management application to obtain a unique index for conceptual row

   creation in the rtpSessionTable.  In this way the SNMP Set operation

   MAY be used to configure a monitor.

Are 'monitors' a mandatory part of the deployment scheme? From previous text I understood that RTCP-XR and the MIB can run in any end-point. What happens there? How are indices correlated in an Internet deployment with thousands of endpoints? How are multiple managers supported?

 

Regards,

Dan

_______________________________________________
Audio/Video Transport Working Group
avt at ietf.org
https://www1.ietf.org/mailman/listinfo/avt