[Battlemesh] any results of the tests?

Elektra onelektra at gmx.net
Mon Apr 2 12:24:19 CEST 2012

Hi -

> In addition to this Madwifi is a dead end that wont allow us to work
> with layer-2 data for routing metrics. Either we get a fixed ath5k or
> we should switch to a better hardware platform.

the old Fonera platform is a dead end, because we have to include 802.11s support in our mesh testbed. Besides, flash access is slow as hell, power consumption is high, CPU performance is low and we have no USB support. We can't exclude 802.11s because the ath5k driver is unstable. 

The Foneras were cheap and small, but they are crappy and have collected a lot of patina. I don't see any reason not to upgrade. We will have to spend some money, but these days alternatives are very affordable.

I guess ath5k is not going to be fixed for the Openwrt "Atheros" port, not now and not in the future. I don't know if Felix is reading this list, but he is the only person I can think of for this task. The incentive to do so should be minimal, since the platform already has stable support with Madwifi. Madwifi includes tons of neat features and it took years to get ad-hoc support stable. It doesn't support things that rely on the Linux 802.11 stack - and that is the bummer in this case. But an excellent excuse to upgrade ;)

I like the idea to buy a bunch of TP-Link 703s - they are compact, AR7XXX based, 32MB RAM, 4MB flash and I estimate you can carry 50 units including the power supplies in a suitcase. As a disadvantage, there is only pcb antennas, so the hack of removing the antenna and shielding the port with tin foil is not possible.

But that idea should not be in the way of someone offering something more powerful with dual radios and 5GHz support...


More information about the Battlemesh mailing list