Re: [Pce] WG Adoption of draft-chen-pce-pcep-ifit-06

Hi Greg,

MNA is quite a new thing. And we cannot predict the solution nor analysis.
For example, I am not sure in this case, if the bSPL is on the top or always at the bottom.
If the bSPL is on the top for Hop by Hop use, I agree there will be packet drop if the node does not support.
It’s also possible to keep the bSPL at the bottom. So the transit nodes do not aware.
My suggestion is to exclude the MPLS encapsulation in this draft until we have a clear NMA solution.
Best,
Tianran
From: Greg Mirsky [mailto:gregimirsky@gmail.com]
Sent: Tuesday, July 5, 2022 3:39 AM
To: Giuseppe Fioccola <giuseppe.fioccola@huawei.com>
Cc: wangaj3@chinatelecom.cn; Dhruv Dhody <dd@dhruvdhody.com>; pce@ietf.org; draft-chen-pce-pcep-ifit@ietf.org
Subject: Re: [Pce] WG Adoption of draft-chen-pce-pcep-ifit-06

Hi Giuseppe,
thank you for your expedient reply. I understand RFC 9197 differently. In my opinion, it applies to nodes that support it, i.e., IOAM, but not all trace options or IOAM data types. It seems reasonable to have MNA extensions in the control and management plane (e.g., YANG data model) advertising MNA capabilities so that an MPLS label stack can be constructed in a manner avoiding MNA bSPL coming to the top on MNA unsupporting node.

Regards,
Greg

On Mon, Jul 4, 2022 at 10:35 AM Giuseppe Fioccola <giuseppe.fioccola@huawei.com<mailto:giuseppe.fioccola@huawei.com>> wrote:
Hi Greg,
Thank you for pointing out the ongoing work on the MPLS Network Actions (MNA).
I agree with you that, for MPLS, this is a possibility to consider especially in case it will conflict with IOAM specification in RFC 9197, where it is stated that a transit node must ignore option types that it does not understand.

Regards,

Giuseppe

From: Greg Mirsky <gregimirsky@gmail.com<mailto:gregimirsky@gmail.com>>
Sent: Monday, July 4, 2022 6:40 PM
To: Giuseppe Fioccola <giuseppe.fioccola@huawei.com<mailto:giuseppe.fioccola@huawei.com>>
Cc: wangaj3@chinatelecom.cn<mailto:wangaj3@chinatelecom.cn>; Dhruv Dhody <dd@dhruvdhody.com<mailto:dd@dhruvdhody.com>>; pce@ietf.org<mailto:pce@ietf.org>; draft-chen-pce-pcep-ifit@ietf.org<mailto:draft-chen-pce-pcep-ifit@ietf.org>
Subject: Re: [Pce] WG Adoption of draft-chen-pce-pcep-ifit-06

Hi Giuseppe,
thank you for your quick response. I agree with your analysis regarding the IOAM and AltMark marked packets in an IPv6 network. And yes, Synonymous Flow Label can be used in an MPLS network to support the AltMark method. But I am not sure that the solution proposed in draft-gandhi-mpls-ioam is the one that is going to be adopted. I would like to note the ongoing work on the MPLS Network Actions (MNA) in the MPLS WG and the IOAM is considered one of the use cases in draft-ietf-mpls-mna-usecases<https://datatracker.ietf.org/doc/draft-ietf-mpls-mna-usecases/>. It might be that IAOM in MPLS would not be supported using G-ACh but MNA with post-stack data (PSD) approach. If that is the case and since MAN will use the newly assigned base Special Purpose Label (bSPL), a node that does not support MNA will drop the packet when that bSPL is discovered as the top label.

Regards,
Greg

On Mon, Jul 4, 2022 at 1:27 AM Giuseppe Fioccola <giuseppe.fioccola@huawei.com<mailto:giuseppe.fioccola@huawei.com>> wrote:
Hi Greg,
Yes, this is the supposed behavior as specified in IOAM and Alt-Mark documents.
For IPv6, IOAM and Alt-Mark are encapsulated in option data fields using extension headers (either HBH or DOH). Both draft-ietf-6man-ipv6-alt-mark and draft-ietf-ippm-ioam-ipv6-options state that nodes that do not support the Option must ignore it, according to the procedures of RFC8200.
For MPLS, it also applies to draft-ietf-mpls-rfc6374-sfl. Looking at draft-gandhi-mpls-ioam, it is also mentioned that the intermediate node that is not capable of supporting the IOAM functions can simply skip the IOAM processing.

Regards,

Giuseppe

From: Greg Mirsky <gregimirsky@gmail.com<mailto:gregimirsky@gmail.com>>
Sent: Saturday, July 2, 2022 4:45 AM
To: Giuseppe Fioccola <giuseppe.fioccola@huawei.com<mailto:giuseppe.fioccola@huawei.com>>
Cc: wangaj3@chinatelecom.cn<mailto:wangaj3@chinatelecom.cn>; Dhruv Dhody <dd@dhruvdhody.com<mailto:dd@dhruvdhody.com>>; pce@ietf.org<mailto:pce@ietf.org>; draft-chen-pce-pcep-ifit@ietf.org<mailto:draft-chen-pce-pcep-ifit@ietf.org>
Subject: Re: [Pce] WG Adoption of draft-chen-pce-pcep-ifit-06

Hi Giuseppe,
I have a question about your statement:
But if nodes on the path do not support some capabilities, it is not a big issue. Indeed, both Alternate Marking and IOAM documents specify that nodes that do not support a specific functionality will forward the packet without any changes to the data fields and they are simply not considered in the measurement.
Is the expectation that a packet marked with IOAM or AltMarking will be forwarded by a non-supporting node applies to all IETF networking technologies, for example in an MPLS network?

Regards,
Greg

On Fri, Jul 1, 2022 at 8:55 AM Giuseppe Fioccola <giuseppe.fioccola=40huawei.com@dmarc.ietf.org<mailto:40huawei.com@dmarc.ietf.org>> wrote:
Hi Aijun,
Thanks for the support.
Regarding your question, I think we can clarify this point in the next version. If a PCE instantiates a path on the PCC with an IFIT capability enabled, it is supposed that there are at least two nodes (e.g. starting and ending node) which support it. But if nodes on the path do not support some capabilities, it is not a big issue. Indeed, both Alternate Marking and IOAM documents specify that nodes that do not support a specific functionality will forward the packet without any changes to the data fields and they are simply not considered in the measurement.

Regards,

Giuseppe

From: wangaj3@chinatelecom.cn<mailto:wangaj3@chinatelecom.cn> <wangaj3@chinatelecom.cn<mailto:wangaj3@chinatelecom.cn>>
Sent: Friday, July 1, 2022 11:26 AM
To: 'Dhruv Dhody' <dd@dhruvdhody.com<mailto:dd@dhruvdhody.com>>; pce@ietf.org<mailto:pce@ietf.org>
Cc: draft-chen-pce-pcep-ifit@ietf.org<mailto:draft-chen-pce-pcep-ifit@ietf.org>
Subject: 答复: WG Adoption of draft-chen-pce-pcep-ifit-06

Hi, All:

I support its adoption.

One questions to the authors:
Is it enough that only the headend support the defined iFIT capabilities? What’s the procedures when the nodes on the LSP/SR path doesn’t support the defined iFIT capabilities?

Aijun Wang
China Telecom

发件人: Dhruv Dhody [mailto:dd@dhruvdhody.com]
发送时间: 2022年6月24日 16:59
收件人: pce@ietf.org<mailto:pce@ietf.org>
抄送: draft-chen-pce-pcep-ifit@ietf.org<mailto:draft-chen-pce-pcep-ifit@ietf.org>
主题: WG Adoption of draft-chen-pce-pcep-ifit-06

Hi WG,

This email begins the WG adoption poll for draft-chen-pce-pcep-ifit-06.

https://datatracker.ietf.org/doc/draft-chen-pce-pcep-ifit/

Should this draft be adopted by the PCE WG? Please state your reasons - Why / Why not? What needs to be fixed before or after adoption? Are you willing to work on this draft? Review comments should be posted to the list.

Please respond by Monday 11th July 2022.

Please be more vocal during WG polls!

Thanks!
Dhruv & Julien
_______________________________________________
Pce mailing list
Pce@ietf.org<mailto:Pce@ietf.org>
https://www.ietf.org/mailman/listinfo/pce