[Battlemesh] Battlemesh v5 tests

Gabriel Kerneis kerneis at pps.jussieu.fr
Tue Mar 6 20:36:43 CET 2012


> Could you be more specific on what you are looking for ?

The algorithm used by BATMAN to choose the route to some node.

> What is missing ?  The B.A.T.M.A.N. routing metric ? How is that related to
> multi-link setups ? Maybe I misunderstand ?

Since BATMAN makes different route choices when it detects multi-link setups
(IIUC), it must do *something* differently.  I assumed that this is done by
tweaking the metrics, but maybe you have another mechanism at work.

I don't really care for the detailed options.  What matters is this mechanism,
this *something* that BATMAN does differently: what is it?  How does it work?  I
failed to find an answer on your wiki.

I am not trolling: I'm genuinely interested in any pointer.  Do you have some
available (even public mailing-list discussions would be a great start)?

> Just for the fun of it I quickly searched for documentation about the famous 
> "babel-z" code/algorithm on the babeld website. Maybe you can help me to find 
> the documentation because I could not find any?!

You must be kidding.  Did you really go to the official babel homepage [1] and
grep for BabelZ?  It gives you the slides [2] and video [3] of a talk about it,
that explain the BabelZ algorithm in much more details than anything I could
find about BATMAN's "alternating interfaces".  That's the kind of material I'm
looking for.

> This makes babel-z a proprietary protocol by your own standards, doesn't it ?

The protocol is indeed not documented, AFAICT.  It does not bother me so much
since the ideas underlying it are explained, which I consider much more

(Note btw. that my standards about what makes something proprietary might differ
from Juliusz'.)

[1] http://www.pps.jussieu.fr/~jch/software/babel/
[2] http://www.pps.jussieu.fr/~jch/software/babel/wbmv4.pdf
[3] http://www.youtube.com/watch?v=1VNzm0shSA8

Best regards,

More information about the Battlemesh mailing list