[Battlemesh] tests for WBMv3

[L. Aaron Kaplan]

On May 26, 2010, at 11:40 AM, Benjamin Henrion wrote:

> On Wed, May 26, 2010 at 11:35 AM, Luca Tavanti
> <luca.tavanti at iet.unipi.it> wrote:
>> Hi Elektra,
>> 
>> Thanks for the quick reply :)
>> 
>> Yes, ping could be a useful tool for generating simple traffic and collecting statistics.
>> 
>> As for running two protocols at the same time, it has pros and cons.
>> You indeed let them work under the same environment, but, what if you want to compare three or more protocols? Set up a sort of tournament?
>> Also, how can you be sure that the protocols do not disturb each other? Especially when you increase the load, this is not such a remote possibility...
>> 
>> Finally, can I ask you how you plan to measure overhead?
> 
> What I did to generate traffic between 2 foneras was to create a 4MB
> file with DD in a ramfs (/tmp/4M), and wget it to /dev/null on the
> otherside.
> 
> I avoid using iperf now, because packet generation takes too much CPU.
> 

Fits to my experience. Especially doing a parallel iperf from many different clients can be a huge CPU hog.

> Otherwise the best is to generate the traffic with computers that are
> not routers.

ACK!