Re: [iccrg] New draft submitted for draft-pan-tsvwg-hpccplus-02.txt

["Pan, Rong" <[e-mail]>]

Michael,

I read the paper that you referenced. Indeed, it is very similar to HPCC++.  Time has finally come to do switch-assisted congestion control 😊. I am not clear about your comment on “path segments”. Data center network should be able to support it end-to-end, right?

Thanks,

Rong

From: Michael Welzl <[e-mail]>
Date: Monday, December 14, 2020 at 11:24 PM
To: "Rui, Miao" <[e-mail]>
Cc: iccrg <[e-mail]>, "Pan, Rong" <[e-mail]>, ""Liu, Hongqiang(洪强)"" <[e-mail]>, "jri.ietf" <[e-mail]>, "Lee, Jeongkeun" <[e-mail]>, Barak Gafni <[e-mail]>, Yuval Shpigelman <[e-mail]>
Subject: Re: [iccrg] New draft submitted for draft-pan-tsvwg-hpccplus-02.txt

Hi !

In this world of research, I do want to comment that the signaling protocol looks extremely similar to what I came up with in my Ph.D. thesis. This paper from 2005 describes it:
Michael Welzl: "Router Aided Congestion Avoidance with Scalable Performance Signalling", KiVS 2005 (Kommunikation in verteilten Systemen), Kaiserslautern, Germany, 28 February - 2 March 2005, Springer-Verlag. https://folk.universitetetioslo.no/michawe/research/publications/kivs2005.pdf
I made this overview page in 2006, still alive:  https://folk.universitetetioslo.no/michawe/research/projects/ptp/index.html

Back then, everyone told me that this is completely undeployable  :-)   but this was before data centers became the big thing they are today, and before there were network cards supporting this kind of thing.

Indeed, such mechanisms are simply no fit for the Internet’s end-to-end congestion control - but they should be deployable on path segments, with PEPs at the edge, like here:  https://www.isi.edu/isi-xcp/docs/kapoor-pep-gi2005.final.pdf

Cheers,
Michael



On Dec 15, 2020, at 1:11 AM, Rui, Miao <[e-mail]<mailto:[e-mail]>> wrote:

Hello ICCRG members,


Alibaba, Intel, and Mellanox have worked on an INT-based High Precision Congestion Control algorithm: HPCC++. We have posted an initial draft that can be found at
https://www.ietf.org/id/draft-pan-tsvwg-hpccplus-02.txt

The key design choice of HPCC++ is to use inband telemetry to provide fine-grained load information, such as queue size and accumulated tx traffic to compute precise flow rates. This has two major benefits:
1. HPCC++ can quickly converge to proper flow rates to highly utilize bandwidth while avoiding congestion;
2. HPCC++ can consistently maintain a close-to-zero queue for low latency.

We would love to hear your comments and feedback.

Best regards,

Rui Miao
_______________________________________________
iccrg mailing list
[e-mail]<mailto:[e-mail]>
https://www.irtf.org/mailman/listinfo/iccrg