Howdy!
When there are two limiting factors when sending data:
* bandwidth of first leg
* bandwidth of slowest leg
Sender application (iperf) can sense bandwidth of first leg as sending
IP stack will buffer certain amount of data (Tx buffer) and then
throttle back the application. In your case it seems that sending
computer's WiFi can transmit at roughly 3Mbps. If that was the overall
bottleneck, receiver wouldn't see any dropped datagrams (or nearly any
of them).
Overall bottleneck (it could be the last leg as well) will impose
dropped datagrams when testing UDP and that will be seen only by
receiving IP stack / application.
Peace!
Mkx
-- perl -e 'print $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'
-- echo 16i[q]sa[ln0=aln100%Pln100/snlbx]sbA0D4D465452snlb xq | dc
------------------------------------------------------------------------
BOFH excuse #9:
doppler effect
Randy Buck je dne 23/03/10 21:46 napisal-a:
Hi all,
I am using Iperf to find the optimal sending rate of my wireless mesh
network. I am running into an issue when I run the following command:
iperf -u -fk -c 5.0.0.1 *-b 5000K* -p 4000 -t 10
I get the following report (or something very similar). I've bolded
parts in question:
------------------------------
------------------------------
Client connecting to 5.0.0.1, UDP port 4000
Sending 1470 byte datagrams
UDP buffer size: 109 KByte (default)
------------------------------------------------------------
[ 3] local 5.0.0.24 port 58708 connected with 5.0.0.1 port 4000
[ 3] 0.0-10.2 sec *3623 KBytes* 2916 Kbits/sec
[ 3] Sent 2524 datagrams
[ 3] Server Report:
[ 3] 0.0-10.4 sec *2703 KBytes* 2128 Kbits/sec 26.951 ms 640/
2523 (25%)
[ 3] 0.0-10.4 sec 1 datagrams received out-of-order
Why is it that when I start the client and specify -b 5000K that I am
sending at a rate much less (3623 KBytes)? Shouldn't Iperf send out
UDP packets as fast as it can? I can understand that the acutal
bandwidth is much less (2703 KBytes) considering I lost 25% of the
packets. Please explain to me why Iperf is not trying to send out at
5000 KBytes. How is the number 3263 KBytes calculated?
Thanks,
Randy Buck
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Iperf-users mailing list
Iperf-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/iperf-users
