Re: Max. MBit/sec ?

[Martin Visser <martinvisser99 () gmail com>]

You can read it from the graph, just look for the highest point ;-)

Alternatively use the "Copy" button and use your awk (or perl or sort) to
find the sample with the maximum value.


Regards, Martin

MartinVisser99 () gmail com


On Fri, Jun 4, 2010 at 12:24 AM, János Löbb <janos.lobb () yale edu> wrote:

> Hi Martin,
>
> I am collecting 540000 packets before I stop the capture.  That is about 3
> minutes, or so.  The .ksh file on Aix 5.3 is this:
>
> #!/bin/ksh
> . $HOME/.profile
> export DHSM=`date +%Y_%m_%d_%H`
> tcpdump -c540000 -ien8 -w/tmp/tcpdump$DHSM.pcap host sequoia and not
> bml0028
>
> the bml0028 machine is my Mac from where I ssh to the sequoia box.
>
> The IO Graph has to do some calculation similar to (bits(t+dt)-bits(t))/dt
>  where t is a given time and dt is a chosen time interval to provide a
> meaningful result.  Of course I can awk the result and collect how many bits
> were transferred back and forth via the pipe in a given time interval - let
> say 1 second - and calculate the "velocity" of the bits at that particilar
> time of the capture, but IO Graph is already doing it, so I do not want to
> reinvent the wheel.
>
> So the question is if I can tap into and get the Maximum bit rate for the
> duration of the capture.
>
> Thanks a lot,
>
> János
>
>
> On Jun 3, 2010, at 9:40 AM, Martin Visser wrote:
>
> Maximum throughput is dependent on the sample time - if you change from 1
> second to 0.1s to 0.01s you will notice higher and higher peaks.
>
> A 1Gbps ethernet pipe is clocked at 1000000000 bits per second so that is
> the maximum rate bits will go. But because of the overhead from headers and
> encapsulation, enforced gaps between packets, the actual thorughput (and
> data goodput) will be different.
>
>  The IO graph is based (I think) simply on the sum of frame lengths in the
> sample period you have chosen.
>
> Regards, Martin
>
> MartinVisser99 () gmail com
>
>
> On Thu, Jun 3, 2010 at 11:29 PM, János Löbb <janos.lobb () yale edu> wrote:
>
> > Hi,
> >
> > Under the Statistics menu there is the Summary menu item displaying the
> > Avg. MBit/sec at the bottom.  Is there a way to find the Maximum MBit/sec
> > value somehow, or find a way to calculate it ?
> >
> > On a 1 Gbit/sec pipe I am getting 130M/bit/sec spikes when I look the I/O
> > diagram of the capture.  I am trying to find out if that pipe is capable for
> > more or not taking into account the network subsystem configuration of the
> > machine.  How I/O Diagram is calculating it ?
> >
> > Thanks ahead,
> >
> > János
> >
> > ___________________________________________________________________________
> > Sent via:    Wireshark-users mailing list <wireshark-users () wireshark org>
> > Archives:    http://www.wireshark.org/lists/wireshark-users
> > Unsubscribe: https://wireshark.org/mailman/options/wireshark-users
> >             mailto:wireshark-users-request () wireshark org
> > ?subject=unsubscribe
>
> ___________________________________________________________________________
> Sent via:    Wireshark-users mailing list <wireshark-users () wireshark org>
> Archives:    http://www.wireshark.org/lists/wireshark-users
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-users
>             mailto:wireshark-users-request () wireshark org
> ?subject=unsubscribe
>
>
>
> ___________________________________________________________________________
> Sent via:    Wireshark-users mailing list <wireshark-users () wireshark org>
> Archives:    http://www.wireshark.org/lists/wireshark-users
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-users
>             mailto:wireshark-users-request () wireshark org
> ?subject=unsubscribe
___________________________________________________________________________
Sent via:    Wireshark-users mailing list <wireshark-users () wireshark org>
Archives:    http://www.wireshark.org/lists/wireshark-users
Unsubscribe: https://wireshark.org/mailman/options/wireshark-users
             mailto:wireshark-users-request () wireshark org?subject=unsubscribe