• Current Meeting Report
• Slides
• Attendees List

Realtime Traffic Flow Measurement (rtfm)

NOTE: This charter is a snapshot of that in effect at the time of the 38th IETF Meeting in Memphis, Tennessee. It may now be out-of-date.

Chair(s): 

Nevil Brownlee <n.brownlee@auckland.ac.nz>
Barry Greene <bgreene@cisco.com>
Sig Handelman <handel@watson.ibm.com>

Operations and Management Area Director(s): 

Scott Bradner <sob@harvard.edu>
Michael O'Dell <mo@uu.net>
Deirdre Kostick <kostick@qsun.att.com>

Area Advisor: 

Mike O'Dell <mo@uu.net>

Mailing Lists: 

General Discussion:rtfm@auckland.ac.nz
To Subscribe: majordomo@auckland.ac.nz
In Body: subscribe rtfm your-email-address
Archive: ftp://ftp.auckland.ac.nz/pub/rtfm/archive

Description of Working Group: 

This working group has three main objectives:

Review existing work in traffic flow measurement, including that of the RMON and Internet Accounting working groups and published work from independent researchers. 
Produce an improved Traffic Flow Model considering at least the following: 
- efficient hardware implementation 
- effect of IPv6 on traffic measurement 
- extension of the accounting model to widen the range of measurable quantities 
- simpler ways to specify flows of interest 
- maintain existing focus on data reduction capabilities 
Develop the Flow Meter MIB as a 'standards track' document with the IETF.

· Real Time Flow Measurement Working Group - New Attributes for Traffic Flow Measurement 

· Traffic Flow Measurement: Meter MIB

Request For Comments:

Current Meeting Report

Minutes of the Realtime Traffic Flow Measurement (RTFM) Working Group 

Reported by: Stephen Stibler and Greg Ruth 

Please note that the Working Group considered its charter carefully and revised the Goals and Milestones - these are included in the minutes. The body of the charter is OK as it is, but the Working Group will consider it on the mailing list and will produce revised text before the Munich meeting.

I. 'New Meter MIB' Internet Draft 

Nevil Brownlee, The University of Auckland, is proposing a new version of the Meter MIB, to improve on RFC 2064 from implementation experience. He gave a presentation explaining the MIB changes proposed in the I-D, and others resulting from his implementation work. The changes are: 

Textual Conventions have been added or extended for the different address types, i.e., MediumType, MediumAddress, PeerType, PeerAddress, AdjacentType and AdjacentAddress. 

The RuleAttributeNumber textual convention has been extended to include MatchingStoD (a 'packet matching' attribute), and to allow the testing of computed attributes. 

A few variables have been added to the Ruleset, Manager and Meter Reader Info Tables so as to provide better control and co-ordination for meters and meter readers. 

A new table of 'flowPackage' objects has been added. This allows a meter reader to retrieve values for a list of attributes from a single flow. The values are returned in a BER-encoded sequence, which provides a simple, standard way to retrieve large amounts of flow data. 

The 'Security Considerations' section has been extended to suggest several ways to protect the meter and its flow data. 

The Working Group agreed that a new version of the draft, with these changes and other editorial changes to the supporting sections, should be published. It will be put to a Working Group last call, then submitted to IESG for publication as a Proposed Standard RFC.

II. Performance Measures for Multimedia Applications 

Randy Bloomfield, Institute for Telecommunication Sciences, Boulder, Colorado (part of NTIA), gave a short presentation on this topic. ITS has focused on developing user-oriented performance measures which are technology-independent, perception-based and objectively evaluated. They have also developed a system to make these measurements by examining the audio/video signal. New measurement techniques like these are required to assess the effects of new impairments to digital audio, video and multimedia data streams, particularly where these streams are carried across a data network. For example: 

· For video: error blocks, tiling, jerkiness, smearing 

· For audio: non-stationary compression artifacts (burbling, wavering, quivering) 

It is important to note that quality often depends on the information complexity at the source. 

This work could provide input for RTFM in suggesting new attributes which are important in measuring transmission of this kind of data. Randy offered a demonstration of the ITS measurement system at a future IETF meeting. He can be contacted at randy@its.bldrdoc.gov. 

III. 'New Attributes for Traffic Flow Measurement' Internet Draft 

Sig Handelman (IBM) presented the latest draft of the I-D. It considers work on performance in other Working Groups in the areas of connectivity, throughput, packet loss and delay metrics; these are also of interest to the BMWG/IPPM, RSVP and RMON2 WGs. In addition, MIB2 has many variable which are of interest in performance measurement. 

The new attributes have three forms: packet traces, aggregates and series. They should be useful for the analysis of performance, congestion and normal versus abnormal network behavior. 

Stephen Stibler (IBM) described his implementation work in this area, showing histograms of forward- and backward- packet inter-arrival times collected for specified UDP flows. Buckets with specified upper bounds were used to group the data within the meter. 

Stephen has not yet attempted to extend the Meter MIB to include the new attributes. The Working Group spent some time discussing how the Traffic Measurement Architecture could be extended to allow specification of: 

· Ways of specifying data set parameters (buckets, histograms, rate-determining intervals, etc.) 

· Complex measures to calculate (e.g., packet loss from TCP analysis) 

· Rule actions required to collect 'new attribute' data from packets 

An area of particular interest is the need to make the architecture extensible so as to allow further new attributes to be added as the need arises. Implementation work on new attributes will also continue, with a view to determining an initial set of the new attributes. The Working Group agreed that these issues should be discussed on the mailing list so as to produce a more complete version of the I-D by August 97.

IV. The RTFM Working Group Charter 

The RTFM Working Group needs to review/revise its charter because: 

· It has moved into a new IETF Area (Transport) 

· There is an increased emphasis on security within IETF

The Working Group discussed its existing charter. It still appears to describe the Working Group's goals well, but needs to make a more definite statement about security. Only one of the original subgoals has not been worked on -developing better ways to specify the flows of interest; this will also be addressed. Discussion of goals and security on the mailing list was requested. 

Our current list of Goals and Milestones was updated to the following: 

April 97: Submit Meter MIB I-D to IESG for publication as Proposed Standard RFC 

August 97: Publish revised and expended I-D on 'New' Attributes Publish revised 

'Architecture' I-D to cover new attributes 

March 97: Submit I-Ds on 'New' Attributes and Architecture as Standards Track RFCs 

This list of goals may change as the Working Group's understanding of the new attributes grows with implementation experience. 

Slides

1. Ruleset Information Table 

2. Basic Meter MIB Compared with RFC 2064 

3. Performance Measures for Multimedia Applications 

Attendees List