From: Daniel S. <dan...@ed...> - 2004-10-29 08:13:28
|
Hi Will, I agree with Ming, such a peak throughput should be possible. Some time I did some tests with a small TCP/IP throughput measurement tool [1] using poor home equipment (a Athlon 1.4 GHz box with RTL8139 network card running Debian as server, a 800 MHz laptop with Windows 2000 running the client via Cygwin). I reached 11.18 MB/sec throughput. So given your equipment (strong CPU's in the client and a NetApp filer as server - this box do commit your writes immediately after putting the data in the NVRAM based write cache), the measured peak should be possible - even when having the additional iSCSI overhead. Daniel [1] = http://www.scheibli.com/v3/projects/stackthru/index.html Ming Zhang said: > it is possible to get a 11.5MB from a 100Mbit network. Do not worry > about this. Just like people get 117MB/s from a 1Gbit network. If you > do duplex, u get more. > > > Ming > > On Thu, 2004-10-28 at 17:41, Will wrote: >> Hi. I am a new IOmeter user and I have come across some interesting >> results. I used IOmeter to test an iSCSI drive on a Network Appliance >> F87 from a single Ethernet client. This was done over a 100Mb link. >> The throughput results peaked at 11.5MB/s, which from my understanding >> is more than is possible/likely for the line speed. Is this expected >> behavior? Is data other than network traffic is included in this >> measurement? Thanks in advance for any insights. >> >> -=wh=- >> >> IOmeter 12.16.03 >> Dual CPU Xeon 3.06 client, 4G RAM. >> IOmeter configured for 64K reads from six disk workers at once. >> Bumped the number of outstanding I/O per target to 16. >> >> >> ------------------------------------------------------- >> This Newsletter Sponsored by: Macrovision >> For reliable Linux application installations, use the industry's >> leading setup authoring tool, InstallShield X. Learn more and evaluate >> >> today. http://clk.atdmt.com/MSI/go/ins0030000001msi/direct/01/ >> _______________________________________________ >> Iometer-user mailing list >> Iom...@li... >> https://lists.sourceforge.net/lists/listinfo/iometer-user > -- > -------------------------------------------------- > | Ming Zhang, PhD. Student > | Dept. of Electrical & Computer Engineering > | College of Engineering > | University of Rhode Island > | Kingston RI. 02881 > | e-mail: mingz at ele.uri.edu > | Tel. (401) 874-2293 > | Fax. (401) 782-6422 > | http://www.ele.uri.edu/~mingz/ > | http://crab.ele.uri.edu/gallery/albums.php > -------------------------------------------------- > > > > > ------------------------------------------------------- > This Newsletter Sponsored by: Macrovision > For reliable Linux application installations, use the industry's > leading setup authoring tool, InstallShield X. Learn more and evaluate > today. http://clk.atdmt.com/MSI/go/ins0030000001msi/direct/01/ > _______________________________________________ > Iometer-user mailing list > Iom...@li... > https://lists.sourceforge.net/lists/listinfo/iometer-user |