[Battlemesh] any (intermediary) results

Linus Lüssing linus.luessing at web.de
Tue Mar 22 14:48:46 UTC 2011


On Tue, Mar 22, 2011 at 02:48:07PM +0100, Marek Lindner wrote:
> On Tuesday 22 March 2011 13:24:32 Pau wrote:
> > The script collected 4 completed tests. One test needs one hour (50*5 pings
> > for each node in the network). I got the latency of each node for each
> > protocol stored in different files.
> 
> It would be nice if we manage to test some real world use cases next time. No 
> real user is interested in pinging the network all the time. Having a fluid 
> youtube experience or VoIP comes to mind. 
> This kind of testing might help to shift focus towards real world needs 
> instead of analyzing whether protocol X was 0.5 seconds faster to find the 
> optimal route than protocol Y.
Hmm, fair point. Simple pinging might not always be representative
for user applications (larger packet sizes, how important is the
latency from the user's point of view, impact of user's
application might have different packet delivery ratio
requirements, ...). However then the question is, how to do fair
comparing of the different protocols then...

Maybe doing something similar to the Mean-Opinion-Score could be a
solution? So for instance give the WBM participants a sheet of
paper and let them use and rate the different protocols,
without them knowing which they are currently testing.

There had been lots of discussions about letting the WBM grow more
or not. I think this could be an awesome thing for the
participants to have fun in actually being involved in the tests
and a good point for letting the event grow, as for these kinds of
tests quite some people would be needed for testing. In the end,
users and developers can share their expriences: Developers get to
know better what the users expect and want and the users might
probably be curious about why one or the other protocol performed
better for this or that application and can directly ask the
developers to have a small insight what might have been going on
behind the curtain. So ultimately, this will bring developers and
users closer together.

The techies could then submit their proposals for the
configurations and the users could propose applications they would
like to get tested.

What do you think about something like this? Is it doable or does it
cost too much effort to "integrate" common users in the tests?
Should the users connect with prepared laptops to the mesh,
running with the routing protocols or should they be able to
connect with any device to the mesh via a mesh router with an AP
interface?

Cheers, Linus

PS: I'd also volunteer to make a first draft of such a question
sheet. I'm so far having general questions of the user experience
as well as application specific questions in mind. I think
something like one or two pages of questions could be a good amount.

> 
> Regards,
> Marek
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
> 


More information about the Battlemesh mailing list