Re: [sip-clf] draft-niccolini-sipclf-ipfix-02

Peter Musgrave <peter.musgrave@magorcorp.com> Tue, 01 June 2010 11:03 UTC

MIME-Version: 1.0
Date: Tue, 01 Jun 2010 07:03:07 -0400
Message-ID: <AANLkTilSs-SQ5qKrGrvpSWIjwOJNRC7Y7Dh0yR3hEXue@mail.gmail.com>
From: Peter Musgrave <peter.musgrave@magorcorp.com>
To: niccolini@nw.neclab.eu, bclaise@cisco.com, brian.trammell@hitachi-eu.com, hkaplan@acmepacket.com, sip-clf@ietf.org
Content-Type: text/plain; charset="ISO-8859-1"
Subject: Re: [sip-clf] draft-niccolini-sipclf-ipfix-02
Precedence: list

Hi IPFIX authors/sipclf:

Some comments on the IPFIX advantages draft:

Introduction
2nd paragraph
- add a reference where IPFIX is first mentioned (is this a reference
to NetFlow 9/RFC3954?) Should the IPFIX architecture be cited
(RFC5470)?

"There are a number of applicable tools"
- is it worth referencing any? Are any open-source?

2.0 NIT:
"These points suggest that, even if today file format has potentially"
add "the IPFIX" after today.

4. Information Model
- should this discussion be incorporated into the problem statement?
(It appears to me the current problem  statement favours 'fine
grained')


Does the draft need to use some of the language from the architecture
doc and discuss the SIP observation point with respect to the NIC and
the SIP stack? Should it discuss possible metering? (As I read 5470
every observation point should have a metering process)
[Or does this stuff belong in the eventual IPFIX-SIPCLF draft]

5.
- I would prefer millisecond level time.
NIT: reference to be updated to WG version of problem statement.

Regards,

Peter Musgrave
(individual)

Re: [sip-clf] draft-niccolini-sipclf-ipfix-02 Peter Musgrave