[Battlemesh] WBMv10: CDF graphs of ping test

[Henning Rogge]

Battlemesh allowed me to finally track down a bug in the address compressor
of olsrd2, which sometimes depending on the order and pattern of addresses
made tlvs vanish... Not sure this helps explaining it, but I consider this
Battlemesh a good success.

Henning

Am 16.06.2017 10:32 schrieb "Juliusz Chroboczek" <
jch at pps.univ-paris-diderot.fr>:

> > BMX7 is the winner in all tests except 8-17 (where it is soundly beaten
> > by Babel).
>
> Looking at the data again, I retract that.
>
>   Babel wins 8-17
>   OLSRv1 wins 8-2
>   9-17 is a draw between BMX7 and OLSRv1
>   9-2 is difficult to judge, I'd say BMX7 wins closely followed by OLSRv1
>
> The surprising result is the rather poor showing of OLSRv2 (was it badly
> configured), as well as the somewhat inconsistent showing of Babel --
> almost perfect results in 8-17, but a fair number of outliers in the other
> tests.  The data is consistent with the time constants for starvation
> avoidance being too large, I'll try reducing them for next time.
>
> -- Juliusz
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20170621/e3c96bf8/attachment.htm>