From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp73.iad3a.emailsrvr.com (smtp73.iad3a.emailsrvr.com [173.203.187.73]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by huchra.bufferbloat.net (Postfix) with ESMTPS id 7C0F021F40B for ; Wed, 15 Oct 2014 15:41:53 -0700 (PDT) Received: from localhost (localhost.localdomain [127.0.0.1]) by smtp26.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id CD0B4803F7; Wed, 15 Oct 2014 18:41:52 -0400 (EDT) X-Virus-Scanned: OK Received: from app60.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp26.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id ADB3A803E7; Wed, 15 Oct 2014 18:41:52 -0400 (EDT) X-Sender-Id: dpreed@reed.com Received: from app60.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by 0.0.0.0:25 (trex/5.2.13); Wed, 15 Oct 2014 22:41:52 GMT Received: from reed.com (localhost.localdomain [127.0.0.1]) by app60.wa-webapps.iad3a (Postfix) with ESMTP id 8D408280060; Wed, 15 Oct 2014 18:41:52 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@reed.com, from: dpreed@reed.com) with HTTP; Wed, 15 Oct 2014 18:41:52 -0400 (EDT) Date: Wed, 15 Oct 2014 18:41:52 -0400 (EDT) From: dpreed@reed.com To: "Wes Felter" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20141015184152000000_29236" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: References: <1412988767.10122173@apps.rackspace.com> X-Auth-ID: dpreed@reed.com Message-ID: <1413412912.57714048@apps.rackspace.com> X-Mailer: webmail7.0 Cc: cerowrt-devel@lists.bufferbloat.net Subject: Re: [Cerowrt-devel] bulk packet transmission X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 15 Oct 2014 22:42:22 -0000 ------=_20141015184152000000_29236 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AI just read the first page of the paper so far, but it sounds like it is= heading in a good direction.=0AIt would be interesting to apply also to ho= me access-point/switches, especially since they are now pushing 1 Gb/sec ov= er the air.=0A =0AI will put it on my very interesting stack.=0A =0A =0A=0A= =0AOn Wednesday, October 15, 2014 3:49pm, "Wes Felter" sai= d:=0A=0A=0A=0A> On 10/10/14, 7:52 PM, dpreed@reed.com wrote:=0A> > The best= approach to dealing with "locking overhead" is to stop thinking=0A> > that= if locks are good, more locking (finer grained locking) is better.=0A> > O= S designers (and Linux designers in particular) are still putting in=0A> > = way too much locking. I deal with this in my day job (we support=0A> > syst= ems with very large numbers of cpus and because of the "fine=0A> > grained"= locking obsession, the parallelized capacity is limited). If=0A> > you do = a thoughtful design of your network code, you don't need lots of=0A> > lock= ing - because TCP/IP streams don't have to interact much - they are=0A> > q= uite independent. But instead OS designers spend all their time=0A> > think= ing about doing "one thing at a time".=0A> =0A> The IX project looks like a= promising step in that direction, although=0A> it still doesn't support su= b-core granularity like Linux does.=0A> =0A> https://www.usenix.org/confere= nce/osdi14/technical-sessions/presentation/belay=0A> =0A> --=0A> Wes Felter= =0A> IBM Research - Austin=0A> =0A> _______________________________________= ________=0A> Cerowrt-devel mailing list=0A> Cerowrt-devel@lists.bufferbloat= .net=0A> https://lists.bufferbloat.net/listinfo/cerowrt-devel=0A> ------=_20141015184152000000_29236 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

I just read the first page= of the paper so far, but it sounds like it is heading in a good direction.=

=0A

It would be interesting to apply also to home acces= s-point/switches, especially since they are now pushing 1 Gb/sec over the a= ir.

=0A

 

=0A

I will put it on m= y very interesting stack.

=0A

 

=0A

 

=0A=0A



On Wednesday, October 15, 2014 3:49pm= , "Wes Felter" <wmf@felter.org> said:

=0A
=0A

> On 10/10/14, 7:52 PM, dpreed@r= eed.com wrote:
> > The best approach to dealing with "locking ov= erhead" is to stop thinking
> > that if locks are good, more loc= king (finer grained locking) is better.
> > OS designers (and Li= nux designers in particular) are still putting in
> > way too mu= ch locking. I deal with this in my day job (we support
> > syste= ms with very large numbers of cpus and because of the "fine
> > = grained" locking obsession, the parallelized capacity is limited). If
= > > you do a thoughtful design of your network code, you don't need l= ots of
> > locking - because TCP/IP streams don't have to intera= ct much - they are
> > quite independent. But instead OS designe= rs spend all their time
> > thinking about doing "one thing at a= time".
>
> The IX project looks like a promising step in = that direction, although
> it still doesn't support sub-core granul= arity like Linux does.
>
> https://www.usenix.org/conferen= ce/osdi14/technical-sessions/presentation/belay
>
> --
> Wes Felter
> IBM Research - Austin
>
> ____= ___________________________________________
> Cerowrt-devel mailing= list
> Cerowrt-devel@lists.bufferbloat.net
> https://lists= .bufferbloat.net/listinfo/cerowrt-devel
>

=0A
 ------=_20141015184152000000_29236--