[Battlemesh] What hardware still works?

[Laurent GUERBY]

On Thu, 2016-02-25 at 00:28 +0100, Juliusz Chroboczek wrote:
> > How much would everyone trust these off-brand Chinese models like the
> > one linked above
> 
> Adam,
> 
> Daniel will correct me if I say something stupid, but I think you're not
> looking at the right hardware.  You don't want .11ac devices, at least not
> yet.  .11ac doesn't bring any range improvements, and only marginal
> improvements in spectral efficiency.  (.11ac improves throughput by using
> up huge amounts of radio spectrum, which is not something you want in
> a dense urban environment.  While MU-MIMO looks interesting, I'm not sure
> it's supported in meshes.)

Hi,

Do you have any relevant data to back "only marginal
improvements in spectral efficiency"?

https://chiliproject.tetaneutral.net/projects/tetaneutral/wiki/NanoBeam

On 40 MHz channels we do about x1.5+ in PC-radio-radio-PC measured TCP
bandwidth between n and ac, both in lab and in real links with Ubiquity
hardware.

Sincerely,

Laurent

> 
> What you want, at least until the OpenWRT driver situation clarifies, is
> a good .11n device with two or three radios and known-good OpenWRT drivers
> (ath9k rocks).  Make sure that the radios are not connected internally
> over something reasonable like PCI or PCIe, not USB or SDIO.  Gigabit
> Ethernet is a welcome plus since you'll doubtless end up either wiring
> some parts of your netowkr or tunnelling over the public Internet.
> 
> (The aging RouterStation Pro rocks.  No, I'm not selling mine.)
> 
> -- Juliusz
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh