[Battlemesh] any (intermediary) results

[Andrew Lunn]

> I totally agree, and we should forget about generating traffic from
> the foneras themselves (such as running iptraf of a webserver on the
> routers themselves).
> 
> We tested iptraf and uhttpd at WBMv2 in Brussels, it was consuming 80% CPU.

We could try beg/borrow/steal/buy maybe 6 wireless devices which have
better performance. These would be used for the "user devices" in
Mareks idea, for VoIP or web browsing. From these devices we can
perform wget of files to represent web browsing and maybe iperf in udp
mode with 128Kbps for voice, and look at the packet loss rate and
maybe latency.

As peers for these wget/iperf i would suggest having two or three
defined exit points from the mesh which are well separated, and then
bring them together using long ethernet cables to one laptop which
acts as http and iperf server. We can then see how well the protocols
pick the exit point from the mesh.

Ideally we can trigger the wget and iperf tests remotely and at the
same time from different "user devices" so we can see how these
streams interact with each other.

        Andrew