tomwaters wrote:
Thanks for the reply miles.
I am not sure what you mean the "10's"...can you explain further.
Also, I appreciate NFS, Samba etc. will have an overhead, but this is in my
view way too low for a gigabit pipe.
How else can I narrow down the issue?
Sorry, I shouldn't have replied the post in a hurry.
112 MBytes/sec (896Mbit/s) looks much better than the results you got
from NFS, Samba. I notice the uperf test you ran lasted for only 10
seconds, it may not reflect the throughput precisely. I suggest you
lengthen the running time of the uperf test. (Let's say 600 seconds.)
And please specify the number of connections, the packet size. All of
these parameters affect the throughput.
If iperf or uperf results turn out good. I'm more suspect the low
throughput you observed is related with the implementation of the
specific application layer protocols such as NFS, Samba.
Thanks,
Miles
_______________________________________________
networking-discuss mailing list
[email protected]
