[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [ipfix] new version of IPFIX per packet information
Dear all,
This memo proposes an excellent approach to use the IPFIX protocol to export performance measurement data. It is clearly in the scope of PSAMP WG and of IPFIX WG.
The draft should propose stable template definitions to export packet information from a meter to a collector, but not only. I t should define a common template to export measure results from the collector to an application. Following I propose a common template 'header' for these 2 aspects:
-- packet information template: Meter to collector -------
flowID
sampleID (packetID)
packetTimestamp
packetLength
--- result template: collector to application ----------
flowID
sampleID
metricID: see IPPM REGISTRY in the RFC Queues
Result
Such measurement technique is clearly respectful of the IPPM framework. The document should have at least an IPPM section to present the metrics measured and discuss each metric.
Regards
Emile
-----Message d'origine-----
De : majordomo listserver [mailto:majordomo@mil.doit.wisc.edu] De la part de Elisa Boschi
Envoyé : mardi 1 mars 2005 21:08
À : ipfix@net.doit.wisc.edu
Objet : [ipfix] new version of IPFIX per packet information
Dear all,
we have submitted a new version of our draft: "export of per-packet
information using IPFIX". The latest version can be found here:
http://www.ietf.org/internet-drafts/draft-pohl-perpktinfo-02.txt
It has been improved taking into consideration the comments we have
received so far and includes a section addressing the management of the
"FlowIDs" we have introduced.
Since it describes an extension to the protocol aimed at providing a
more efficient way to export packet information (without requiring any
modification to IPFIX), we propose to merge our draft with the IPFIX
protocol draft. (find our suggestion at the end of this email)
We would like to hear the opinion of the group about this. Since the
export of per-packet information is a relevant issue for PSAMP, and all
IPFIX doocuments are currently in last call, merging the current
proposal with a PSAMP protocol draft is, to our opinion, a valide
alternative.
In both cases the FlowID field should be added in the IPFIX information
model (as a numeric identifier for the flow)
The IPFIX prootocol draft could be modified as follow:
> Table of Contents
>
> 1. Points of Discussion.........................................4
> 2. Introduction.................................................5
> 3. Terminology..................................................6
> 4. Criteria for Flow Expiration and Export.....................12
> 5. Message Format..............................................13
> 6. IPFIX Message Format........................................15
> 7. Specific Reporting Requirements.............................26
One section could be added here describing the special case of export of
packet information (the section could eventually be between 6 and 7 too.
> 8. "Export Time" Computation and Flow Record Time..............29
> 9. Linkage with the Information Model..........................31
> 10. Variable Length Information Element........................32
> 11. Template Management........................................33
> 12. The Collecting Process's Side..............................36
> 13. Transport Protocol.........................................38
> 14. Security Considerations....................................47
> 15. IANA Considerations........................................51
> 16. Examples...................................................52
Add here the example in the last section of our draft.
> 17. References.................................................58
> 18. Acknowledgments............................................6
Regards,
Elisa
--
Help mailto:majordomo@net.doit.wisc.edu and say "help" in message body
Unsubscribe mailto:majordomo@net.doit.wisc.edu and say
"unsubscribe ipfix" in message body
Archive http://ipfix.doit.wisc.edu/archive/
--
to unsubscribe send a message to psamp-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/psamp/>