[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <9B2F6EEB-BD3D-4216-8481-13BAF0D29B9C@netapp.com>
Date: Tue, 13 May 2014 09:40:19 +0000
From: "Eggert, Lars" <lars@...app.com>
To: Hagen Paul Pfeifer <hagen@...u.net>
CC: Stephen Hemminger <stephen@...workplumber.org>,
Cong Wang <cwang@...pensource.com>,
Florian Westphal <fw@...len.de>,
netdev <netdev@...r.kernel.org>,
"sbens@...rosoft.com" <sbens@...rosoft.com>,
"Dave Thaler" <dthaler@...rosoft.com>
Subject: Re: [next PATCH 0/5] net: tcp: DCTCP congestion control algorithm
Hi,
On 2014-5-13, at 11:34, Hagen Paul Pfeifer <hagen@...u.net> wrote:
> Some IETF background: IETF Meeting 79 (Beijing) Microsoft employees
> presented DCTCP with the intention to standardize DCTCP. The rough
> consensus was the DCTCP is datacenter specific, thus closed door
> technique and in the result not of broader interest for TCPM working
> group. Further effort slowed down quickly.
>
> Last IETF meeting in London (IETF 89) Lars Eggert (cc'ed him) picked
> up DCTCP again. My impression was that the willing to adopt or work on
> DCTCP was limited (the way I see it) but not definitely buried.
I presented for Stephen Bensley from Microsoft (CC'ed), who is the main driver of the IETF spec. (My contribution was mostly having an intern do a FreeBSD implementation, which is being reviewed for inclusion over there.)
While the TCPM working group may or may not decide to take this spec on as a work item, there is also a newly starting "datacenter latency control (DCLC)" research group on the IRTF side that may give this a home if TCPM takes a pass. (Mailing list https://www.irtf.org/mailman/listinfo/dclc; wiki http://wiki.tools.ietf.org/group/irtf/trac/wiki/dclc)
Lars
> The IPR statement licence it under Royalty-Free, Reasonable and
> Non-Discriminatory licence.
>
> Hagen
Download attachment "signature.asc" of type "application/pgp-signature" (274 bytes)
Powered by blists - more mailing lists