Re: wireshark and RTT

[Martin Visser <martinvisser99 () gmail com>]

Simone,

Way back about 4 years ago, when I was doing an application response time
assessment, I also found some inconsistencies in the way that Wireshark
sometimes calculated RTT. The main issue was that it didn't always
correctly associate the appropriate ACK with right SEQ. I'd probably want
to look at the specifics of what you are seeing to see if it was the same
problem I was observing. For what it is worth, my post from 2009 is here :-
http://www.wireshark.org/lists/wireshark-users/200906/msg00068.html

Martin

Regards, Martin

MartinVisser99 () gmail com


On 27 August 2013 17:52, Simone Ferlin-Oliveira <ferlin () simula no> wrote:

> Hi,
> I am capturing traces at client and server and only transferring in
> downlink, download (s->c) direction.
> I am plotting different things about these flows - window size, rtt, etc.
> Depending on the plot, either one trace file or the other is appropriate,
> but even like this, wireshark shows sometimes very uncommon RTT values. I
> compare them with tcptrace to validate. I know approximately in which range
> they should be and
> sometimes wireshark shows different values compared to tcptrace -Z ... . I
> am relying on tcptrace for now. captcp was another
> option, but since I capture only 150B for each packet. The tool has a
> small bug because it accepts only full packets. I modified
> the code, but the calculations are wrong since it considers full packet
> sizes only...
>
>
>
> On 26 August 2013 15:37, NITIN GOYAL <nitinkumgoyal () gmail com> wrote:
>
> > On Mon, Aug 26, 2013 at 6:53 PM, Simone Ferlin-Oliveira <ferlin () simula no
> > > wrote:
> >
> > > Hi, I am doing a download from server to client and I want to see the
> > > RTT estimates.
> > > Why is that so complicated? :)
> > > Thanks for the link, but there is no conclusion there, right? What was
> > > the conclusion of your discussion with the wireshark forum people?
> >
> > I did not got conclusion as I was capturing at the client end and it
> > didnot have the access to the server to take the reverse values to validate.
> >
> > It is not complicated but it is like how exactly you are taking the
> > captures. IF you say the RTT value is not convincing, how you say they are
> > not? Actually there are other ways/tools to validate the values.
> >
> > Like you can capture the trace and put the trace with the help of
> > tcptrace tool which also calculate the RTT.
> >
> > Also, you can take reverse RTT calculation frm the server and validate if
> > that will work for you.
> >
> > For me, i did like that with the help of tcptrace i was getting a bt
> > different values then showed by wireshark which looked convincing to me.
> >
> > Regards
> > Nitin
> >
> >
> > ___________________________________________________________________________
> > 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