[Battlemesh] recommendations for 40Km Link / OpenWRT on ubnt ath9k hw ? / AirMax

Ben West ben at gowasabi.net
Tue Mar 25 17:09:24 CET 2014


When I inquired on this list previously about strategies for dealing with a
mix of strong / weak clients for PtMP, someone offered this pointer for
JaldMAC, an 802.11 polling implementation in ath9k:

http://matthias.vallentin.net/papers/nsdr10.pdf
https://github.com/shaddi/jaldimac/commits/master

Unfortunately, that repo has sat idle for 3 years, and I believe it it was
more a proof of concept rather than a usable implementation.



On Tue, Mar 25, 2014 at 9:08 AM, Musti <musti at wlan-si.net> wrote:

> Exactly. Anyone working on implementing a time-slot algorithm for ath9k,
> that would be a very welcome addition. 5GHz band is generally used for
> longer distances and anything that is multipoint at this range is very
> likely to have a hidden node problem.
>
> We have tested ad-hoc at 5GHz as well, in the previously described
> situation, but it performed even worse.
>
> Kind regards,
> Musti
>
>
>
> On 25 Mar 2014, at 14:53, fboehm <fboehm at aon.at> wrote:
>
> > Sounds like a perfect example for a hidden-node(s) situation. It totally
> makes sense that AirMax is faster in that case. Every other MAC Layer based
> on polling clients or allocating timeslots would be better than plain
> CSMA/CA.
> >
> > Regards,
> > Franz
> >
> > Am 25.03.14 14:35, schrieb Musti:
> >> The setup we had running with AirMax for about a year:
> >>
> >> AP on a mountain, about 5 clients around in 20km radius. With AirMax
> throughput was about 10times what we get now with openwrt, with
> significantly more packet loss. Prior to that we found out that AirOS
> without AirMax is comparable to openwrt.
> >>
> >> Regards,
> >> Musti
> >>
> >>
> >> On 25 Mar 2014, at 14:18, Laurent GUERBY <laurent at guerby.net> wrote:
> >>
> >>> On Tue, 2014-03-25 at 13:52 +0100, Musti wrote:
> >>>> Hi,
> >>>>
> >>>> in wlan slovenia network we are slowly looking at migrating from AirOS
> >>>> to openwrt for UBNT devices, however the tests for now show that ath9k
> >>>> performs worse then UBNT AirMax. The difference is rather small in PtP
> >>>> links, however fails horribly at PtM as there is no TDMA.
> >>>
> >>> Hi,
> >>>
> >>> We tested with and without AirMax and found out not much differences
> >>> except in certain case (colocation of close antennas) AirMax
> >>> wasn't working at all (horrible bandwidth performance, slow to access
> >>> web admin). In the end we've been running AirOS without AirMax for a
> >>> while now both PtP and PtM
> >>>
> >>> Do you have detailed performance results with and without AirMax ?
> >>>
> >>> Sincerely,
> >>>
> >>> Laurent
> >>>
> >>> _______________________________________________
> >>> Battlemesh mailing list
> >>> Battlemesh at ml.ninux.org
> >>> http://ml.ninux.org/mailman/listinfo/battlemesh
> >>
> >>
> >>
> >> _______________________________________________
> >> Battlemesh mailing list
> >> Battlemesh at ml.ninux.org
> >> http://ml.ninux.org/mailman/listinfo/battlemesh
> >>
> >
> > _______________________________________________
> > Battlemesh mailing list
> > Battlemesh at ml.ninux.org
> > http://ml.ninux.org/mailman/listinfo/battlemesh
>
>
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
>
>


-- 
Ben West
http://gowasabi.net
ben at gowasabi.net
314-246-9434
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20140325/2effeb7f/attachment-0001.html>


More information about the Battlemesh mailing list