Date: Tue, 30 May 2006 10:44:43 -0700 From: Brooks Davis <brooks@one-eyed-alien.net> To: Robert Watson <rwatson@FreeBSD.org> Cc: Paul Allen <nospam@ugcs.caltech.edu>, "current@freebsd.org" <current@FreeBSD.org>, gallatin@FreeBSD.org Subject: Re: Importing iSCSI target from NetBSD Message-ID: <20060530174443.GC6706@odin.ac.hmc.edu> In-Reply-To: <20060530094413.W79162@fledge.watson.org> References: <447AB34C.4030509@sippysoft.com> <11410450515.20060529225555@lacave.net> <447B77AF.9060309@samsco.org> <447B7A55.7040704@FreeBSD.org> <447B7CB7.5000000@FreeBSD.org> <447B8900.4050603@samsco.org> <20060530004328.GF28128@groat.ugcs.caltech.edu> <20060530015234.GB26022@odin.ac.hmc.edu> <20060530094413.W79162@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--hYooF8G/hrfVAmum Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, May 30, 2006 at 09:46:39AM +0100, Robert Watson wrote: >=20 > On Mon, 29 May 2006, Brooks Davis wrote: >=20 > >On Mon, May 29, 2006 at 05:43:28PM -0700, Paul Allen wrote: > >>>From Scott Long <scottl@samsco.org>, Mon, May 29, 2006 at 05:51:28PM= =20 > >>>-0600: > >>>>P.S. Just to make it clear - just consider running iSCSI over 100MBps > >>>>link or even a slower WAN links, which I think covers very large mark= et > >>>>for this technology now. Performance constrain imposed by running in > >>>>userland is unlikely to be an issue at all. > >>> > >>>Every company and group that I've talked to about iSCSI is worried abo= ut=20 > >>>performance. In any case, please follow the lead of Mr. Senault and= =20 > >>>look at making this a port. > >> > >>And in particular the anticipation of low(er) cost 10Gb Ethernet is a= =20 > >>driving factor behind iSCSI. > >> > >>AFAIK, the low-latency performer in this field (for NICs) is from=20 > >>Myricom. Andrew Gallatin (one of the FreeBSD alpha committers) was=20 > >>responsible for porting the myrinet drivers, so perhaps he can comment = as=20 > >>to whether FreeBSD will be getting a driver for their 10GbE cards. =20 > >>Ethernet at these speeds is real stress-test for many OSs; it should be= =20 > >>interesting to see how FreeBSD holds-up. > > > >There's a driver in current. We don't perform nearly as well as we shou= ld=20 > >at the moment. >=20 > FYI, I recently received donated hardware from Yahoo! and Drew has kindly= =20 > offered to send me a couple of 10gbps cards to work with, so I hope to ha= ve=20 > a chance to start doing some measurement and optimization work. One of t= he=20 > problems we've been having is that it's hard to optimize the CPU use of t= he=20 > network stack when the CPU significantly outstrips available bus and=20 > network bandwidth. It seems like hardware swings back and forth quite a= =20 > bit -- for a few years gigabit was way-the-heck-faster-than-CPU, now it's= =20 > the other way around again. The best stack optimization work happens whe= n=20 > you have to figure out how to get the network stack to perform well in=20 > near-infinite bandwidth scenarios with a CPU-bound stack, which is where = we=20 > are with 10gbps currently. One of the things that makes all this rather= =20 > tricky is that it's quite hard to build test rigs, test setups, and get t= he=20 > hardware details right. Hopefully, with Yahoo's and Drew's help, my test= =20 > setup will be good for looking at this for a couple of years. That's great news. This problem of relative performance leapfrogging is a pain. -- Brooks --=20 Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 --hYooF8G/hrfVAmum Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQFEfIOaXY6L6fI4GtQRAtH+AKDArTb/JGxLZXPCZEXxUeKB0fuwjQCdHUur 8Fk7K7tm7N6T57SoLDpeHpo= =QSyz -----END PGP SIGNATURE----- --hYooF8G/hrfVAmum--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060530174443.GC6706>