Re: [ippm] Request for reviews of OWAMP/TWAMP/STAMP on LAG

["[e-mail]" <[e-mail]>]

Hello Haoyu, Tianran and all,

Thank you for your review and discussion. Some information added with prefix [LZQ].



Best Regards,
Zhenqiang Li

[e-mail]
 
From: Tianran Zhou
Date: 2023-04-28 07:59
To: Haoyu Song; Marcus Ihlar; IETF IPPM WG ([e-mail])
Subject: Re: [ippm] Request for reviews of OWAMP/TWAMP/STAMP on LAG
Hi Haoyu,
 
Thanks very much for your review and questions.
Please see inline for the clarification.
 
Best,
Tianran
 
From: ippm [mailto:[e-mail]] On Behalf Of Haoyu Song
Sent: Wednesday, April 26, 2023 1:43 AM
To: Marcus Ihlar <marcus.ihlar=[e-mail]>; IETF IPPM WG ([e-mail]) <[e-mail]>
Subject: Re: [ippm] Request for reviews of OWAMP/TWAMP/STAMP on LAG
 
IPPM,
 
I have read the documents and have several questions to the authors for clarification.
The document says that the measurement results can be used for traffic steering. AFAIK, LAG uses hashing for load balancing. Are you suggesting the new methods/algorithms for traffic distribution? Is there any existing or planed work for this? 
ZTR> Yes, LAG normally use hash for LB. But as for the described use case, there is requirement to explicitly control the traffic steering. We already implement this, but I do not think more standard work is necessary.
 
[LZQ]Please refer RFC8668 Advertising Layer 2 Bundle Member in ISIS. RFC9085 BGP LS Extensions for Segment Routing, section 2.2.3. And some drafts. Related work is done or in progress.
 
This seems a proposal to use L3 protocols for L2 measurements. Is the L2 feature visible and manageable to the L3 entity? 
ZTR> Maybe it’s not L3 protocol, but actually L4, anyway not relevant J. That’s why we need an identifier, i.e., micro session id, to associate the L2 interface with the test session. So that the test session (L4), know the link between two L2 interfaces.

[LZQ]The measurement results are reported to SDN controller or NMS by telemetry. Then the real time performance of the whole network including the layer 2 link bundle is visible on SDN controller or NMS.

Can you list the performance metrics that are measured and of interest to the LAG case for readers to better understand the motivation?
ZTR> I am not sure if I can list all. Because it depends on the specific use cases, although in theory all metrics could be possible . It may be delay or loss in most case. I think we can mention the metrics used in the described traffic steering use case in this draft.  

 [LZQ]The metrics such as delay, jitter, packet loss etc that apply to OWAMP/TWAMP/STAMP apply to micro OWAMP/TWAMP/STAMP defined in these drafts .

Thanks!
 
Best,
Haoyu
 
From: ippm <[e-mail]> On Behalf Of Marcus Ihlar
Sent: Wednesday, April 19, 2023 8:08 AM
To: IETF IPPM WG ([e-mail]) <[e-mail]>
Subject: [ippm] Request for reviews of OWAMP/TWAMP/STAMP on LAG
 
Hi IPPM,
During the IPPM session at IETF 116 the authors of  draft-ietf-ippm-otwamp-on-lag and draft-ietf-ippm-stamp-on-lag expressed that they believe the documents are ready for WGLC.
However, since these documents were adopted quite recently we’d like to ask the WG to take some time to read them and provide feedback.
 
In response to this mail, please indicate whether you’ve read the document along with any feedback you might have.
 
Note that this is not an official working group last call. 
 
The documents can be found here:
https://datatracker.ietf.org/doc/draft-ietf-ippm-stamp-on-lag/
https://datatracker.ietf.org/doc/draft-ietf-ippm-otwamp-on-lag/
 
HTML versions:
https://www.ietf.org/archive/id/draft-ietf-ippm-stamp-on-lag-01.html
https://www.ietf.org/archive/id/draft-ietf-ippm-otwamp-on-lag-01.html
 
BR
Marcus