[Battlemesh] WBMv6 test discussion [was: Re: Recon Team - Follow Up]

[Simon Wunderlich]

Hey Clauz,

On Fri, Oct 26, 2012 at 12:26:13PM +0100, Clauz wrote:
> Hello!
> 
> On 10/25/2012 10:33 PM, Frank Fitzek wrote:
> > On 24.10.2012 20:47, Simon Wunderlich wrote:
> >> Some tests have been defined for WBMv5 here:
> >> http://battlemesh.org/BattleMeshV5/Tests
> > Awesome. These test topologies are great. We just need to refine the
> > parameters and we are done.
> 
> Well, thanks a lot :)
> 
> I was in charge of the tests in WBMv5, but at the end we didn't perform
> any of them. The main reason is that the devices were ready to deploy
> only in the very last hours of the event. So I suggest, if possible,
> speeding up the firmware setup and flashing process. Perhaps part of it
> can be done before the event?

If we find someone who wants to prepare the firmware for the battle
(still an open position on the list), I'm sure it is no problem to send
him/her some hardware to check/optimize/speed up flashing and try out
the firmware and drivers before the event. 

I think this would be a very important step if we want to actually test
something this time and not debug wifi/routers. :) We have quite some
skilled and experienced OpenWRT hacker on the participant list last year,
I still hope that one of them volunteers for the job. ;)


> 
> The criterion that I tried to use is to replicate as much as possible
> what happens in real-world community networks (for example opted for
> complete shutdown of nodes as opposed to single interfaces going down).
> 
> And, one decision that who is in charge of the tests has to face is
> whether to run all the protocols in sequence or at the same time. Both
> modalities have advantages and drawbacks.

When running the protocols at the same time, temporary effects will affect
all protocols in the same way, but maybe there is also interference between
the protocols ... but some of the protocols require running the protocol
tests separately anyway (like the "11 minutes downloads/21 minutes download"
tests).

BTW, what about mobile test? Maybe we can do tests with robots or parrot ar drones?
:) I don't know if we get very "smart" tests with that (in sense of well
defined, repeatable, ...), but this might be fun too.

Cheers,
	Simon
> 
> ciao,
> Clauz
> 
> 
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20121029/fe0eaf9a/attachment.pgp>