Re: [lwip-users] Help regarding packet arrival times.

[N V Krishna]

On Thu, 30 Sep 2004, K.J. Mansley wrote:

#On Thu, 2004-09-30 at 09:33, N V Krishna wrote:
#> On Thu, 30 Sep 2004, K.J. Mansley wrote:
#>
#> #On Thu, 2004-09-30 at 09:22, N V Krishna wrote:
#> #> Hi,
#> #> I am trying to validate tcp servers against denial of service attacks by
#> #> model checking. To validate this, I want to tally my tool with the
#> #> simulated tcp server. I am using simhost + lwip for this.
#> #>
#> #> For this, I need to know the exact time at which simhost sees a incoming
#> #> packet for the first time. Could anyone please give me any pointers? I
#> #> need this to get a correct estimate of the exact packet inter arrival time
#> #> as seen by the server.
#> #
#> #What operating system are you using?  Would a packet capture and
#> #analysis tool such as ethereal be of any use?
#>
#> I am using it on Linux. I am using Spak to send packets. But I want to
#> know the timings of packets as seen by simhost.
#
#If nothing external will do, you'll have to alter the source to measure
#the time internally.  Measuring time on linux is quite easy though, so
#shouldn't be too hard.  e.g. gettimeofday() each time a packet is
#received.  There will be some "probe effect" (i.e. the measurements will
#take some time, and so may affect the timings you trying to measure)
#associated with this though.


Exactly. That's what I am doing -- I am trying to add code to measure the
time. I am not too well versed with the code. [ I am getting to know it
better, every day. ] I wanted some pointers to help me find a place in the
code to place the probes. Yes, I am aware of the effect the "probe" will
have on the timing. I am ready to test that.

Thanks for your time and reply.

Krishna.