Hi Thomas,
thank you for the feedback and proposed update. Please find my notes below
tagged by GIM2>>.

Regards,
Greg

On Fri, Aug 18, 2023 at 6:56 AM <[email protected]> wrote:

> Dear Greg,
>
>
>
> Thanks a lot for addressing my comments.
>
>
>
> GIM> It could be easier to take out "flow" altogether. WDYT?
>
>
>
> TG> Let me propose something else:
>
>
>
> Change from
>
>
>
> "When analyzing the availability metrics of a service flow between two
> nodes"
>
>
>
> To
>
>
>
> "When analyzing the availability metrics of a connectivity service between
> two measurement points"
>
GIM2>> Prior to IETF-117 the authors extensively discussed the definition
of a connectivity service. Because we couldn't find it being formulated in
published documents we agreed to avoid referencing it in the PAM document.
I hope that you will agree to the following update:
OLD TEXT:
   When analyzing the availability metrics of a service flow between two
   nodes, a time interval as the unit of PAM needs to be selected.
NEW TEXT:
   When analyzing the availability metrics of a service between two
   measurement points, a time interval as the unit of PAM needs to be
   selected.

>
>
> GIM>> I agree. Check the attached diff.
>
>
>
> TG> Perfect thanks!
>
GIM2>> Great!

>
>
> Best wishes
>
> Thomas
>
>
>
> *From:* Greg Mirsky <[email protected]>
> *Sent:* Monday, August 14, 2023 10:33 PM
> *To:* Graf Thomas, INI-NET-VNC-HCS <[email protected]>
> *Cc:* [email protected]; [email protected]
> *Subject:* Re: [OPSAWG] draft-ietf-ippm-pam-04,
> draft-clemm-opsawg-pam-ipfix-00
>
>
>
> Hi Thomas,
>
> thank you for supporting this work and for your helpful comments. Please
> find my notes inlined below tagged by GIM>>.
>
>
>
> Regards,
>
> Greg
>
>
>
> On Wed, Jul 26, 2023 at 2:24 PM <[email protected]> wrote:
>
> Dear Alex and Greg,
>
>
>
> I reviewed draft-ietf-ippm-pam-04 and draft-clemm-opsawg-pam-ipfix-00 and
> have some comments and questions.
>
>
>
> Section 3.1 of draft-ietf-ippm-pam-04 (
> https://datatracker.ietf.org/doc/html/draft-ietf-ippm-pam-04#section-3.1)
> mentions the term "service flow".
>
>
>
> I haven't been able to find in any IETF document describing/defining the
> term. I suggest to describe it in the terminology section 2.1 and specify
> it as an IPFIX and YANG element.
>
> GIM>> I checked and found that "service flow" is used only once in the
> document:
>
>
>
>    When analyzing the availability metrics of a service flow between two
>
>    nodes, a time interval as the unit of PAM needs to be selected.
>
>
>
> It could be easier to take out "flow" altogether. WDYT?
>
>
>
> Section 3.3 of draft-ietf-ippm-pam-04 specifies an "Unavailability
> threshold". I suggest to specify in Section 3.1 of
> draft-clemm-opsawg-pam-ipfix-00 (
> https://datatracker.ietf.org/doc/html/draft-clemm-opsawg-pam-ipfix-00#section-3.1)
> this as IPFIX element as well.
>
>
>
> The "service flow" describes that the SLO metrics are measured between two
> nodes. However in draft-clemm-opsawg-pam-ipfix the SLO metrics are measured
> and export on one node. I would appreciate if you could describe how this
> information is being measured. Presumably by leveraging of probing.
>
> GIM>> The PAM document will not specify how to measure but where and what
> to be measured. We will work on clarifying that all PAM metrics are
> measured between two Measurement Points in the PAM IPFIX draft.
>
>
>
> In general I feel that draft-ietf-ippm-pam-04 would benefit of having a
> reference to the Performance Metrics Registry defined in RFC 8911/8912.
>
>  GIM>> I agree. Check the attached diff.
>
>
>
> I would be interested to understand wherever the authors intend to create
> a draft document describing a service YANG module.
>
> GIM>> That's a great suggestion, thank you. Let us discuss it.
>
>
>
> Best wishes
>
> Thomas
>
> _______________________________________________
> OPSAWG mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/opsawg
>
>
_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to