Hi,
On Dec 31, 2008, at 8:07 PM, Adrian Farrel wrote:
That is my fault :-(
The I-D is at revision -03
http://www.ietf.org/internet-drafts/draft-ietf-pce-monitoring-03.txt
You still have until Monday next week if you want to make comments.
As my punishment, I have gone through Fabien's email to check which
of his coments still apply...
Below my comments on monitoring draft:
Section 3.1:
Example 3: The request must also contain a MONITORING object.
Issue stands.
That was "in addition to" but the text had been clarified, thanks.
Section 3.2:
The format of a PCReq message is as follows:
Must be
The format of a PCMonRep message is as follows:
Issue stands
Fixed.
Section 4.1:
Seems the first sentence is repeated twice.
Fixed in revision -03
Besides it is said that the Monitoring object MAY be included in
PCReq and PCRep so I think it would be good to add the enhanced
PCReq and PCRep BNF description.
Issue stands
Added.
Also in 4.1
s/PCEReq/PCReq/
Fixed.
General comment:
The in-band monitoring request is not fully clear to me.
You are right that this is not clearly explained in the current text.
If there are several path computation requests in the PCReq
which one is targeting?
i.e. Must the PCE return the processing time for each request
seperatly
or a single processing time for all requests?
We could decide to have one MONITORING object per request in the
bundle case but this does
introduce some complexity and is not likely to be that useful. I'd
rather propose to keep the current
mode with one MONITORING object per PCMonReq or PCReq message. When
present the
monitoring request applies to all requests of the bundle. Text added
in section 3.2.
"
As pointed out in [PCEP] we can have several situations:
o Bundle of a set of independent and non-synchronized path
computation requests,
o Bundle of a set of independent and synchronized path computation
requests (SVEC object defined below required),
o Bundle of a set of dependent and synchronized path computation
requests (SVEC object defined below required).
In the case of a bundle of a set of request, the MONITORING object
SHOULD only be present in the
first PCReq or PCMonReq message and the monitoring request applies to
all the request of the bundle,
even in the case of dependent and/or synchronized requests sent using
more than one PCReq or
PCMonReq message.
In first case what happens if there is an SVEC object and Path
Requests
are spread in multiple PCRep messages. Must the Monitoring object be
repeated?
See above.
If a PCRep contains a Monitoring object, is it possible that
this PCRep contains Path replies not related to the Monitoring
request?
Yes because in the BNF, the <metric-pce-list> is per in the <response>
Thanks for the comments.
Cheers,
JP.
Issue stands.
Happy New Year
Adrian
_______________________________________________
Pce mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pce
_______________________________________________
Pce mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pce
