[Battlemesh] configuration files

[Markus Kittenberger]

>
> I don't think it makes sense NOT to use the automatic rate selection.
> Its even essential for any routing protocol that takes the link speed
> into account.
>
> Tests without this algorithm would be quite unrealistic in my opinion.
>

depends, some network still use fixed bitrates,..
and to some extend this imho makes stil sense )-;

but imho battle mesh tests should lean towards recent/future
configurations, not legacy ones,..


> Henning Rogge
>



> > Tests without this algorithm would be quite unrealistic in my opinion.
>
> So how do you know at which bitrate the radio is on?
>
you don`t as the radio only has a bitrate if u manually set one *G

else it has usually just some ideas of which bitrates to use/try per
neighbour,.. (-;

>
> Before testing with variable bitrate, you should fix it so that it is
> a parameter out of the equation.
>
> If you have it variable, you should log its value through time.
>

infact we would need to log bitrate usage patterns per neighbour (as its
usually more than one bitrate that gets  used per any typical logging
interval)

<zynic>
and somehow i doubt that someone really wants to look at this logs
afterwards,.

so hmm, maybe we should test just with fixed bitrate
</zynic>

btw how many protocols in test have bitrate aware metrics?

>
> And there is the story of the multicast bitrate as well, which is
> lower then the transmit bitrate.
>
depends (-;

Markus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20120329/04225373/attachment-0001.html>