[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ns] Need to measure delay



Hi!

I am also measuring end-to-end delays...
For my simulations it wasn't feasible to post-process
my tracefiles, because they were around 700 Megs large and I
had to do plenty of simulations...

Instead I inserted a new "sendtime" field into the common-
header, where I filled in the sendtime. Upon receive, I could
calculate the difference/delay for that packet, means of all delays etc...

Somehow I couldn't use the existing timestamp, because it was used and
manipulated for other purposes (queue-delay measurement)

I know, that inserting a new field into the a separate header would
be nicer, but using a common header field was the easiest way I could
think of...

I someone knows a better way, post it! I'm also very interested!

Bye
Felix

JORGE MANRIQUE MARGALEJO wrote:

> Hi all,
>
> I'm a Spanish student working on DiffServ. I have started to use NS and it seems to be quite adequate for my needs. Nam is a good tool to see whether the network you have designed is working as you thought it would. However, I need to collet statistics on packet loss, and, most important, packet delay and make comparisons among different flows. I've searched through NS manual but surprisingly there doesn't seem to be a solution for that.
>
> I'm thinking about writing on C some kind of post-processing function for the trace files or working with macros on Excel. However I'm sure I'm not the first person who has needed to measure a delay with NS so I thought there must be some existing tools to do that.
>
> Srihari Raghavan mentioned some usage examples to do that as an answer to a question posted by Andreas Huber. Are those examples already accessible, Shihari? Thanks a lot.
>
> Can anybody else help me? I'd be eternally grateful to anyone who could help me, as measuring delays is a MUST for the project I'm working on.
>
> Thanks in advance,
>
> Jorge Manrique