[Make-wifi-fast] [bufferbloat-fcc-discuss] [Cerowrt-devel] arstechnica confirms tp-link router lockdown

[Dave Taht]

It's a really big present and future market, and ideas like these are
being explored. Multiple wifi-router-with-display projects were
launched on kickstarter, some are shipping. I evaluated one, only to
find it was all gui, slapped on top of an antiquated kernel. Other
enterprise players focused on making hundreds of devices easier to
manage, without focusing on the underlying tech much.

Currently hot is the idea of using an app to control the device,
rather than a web browser. eero is doing this, so are others.

but my own thing is this tiny, tiny little bit of an overall next-gen
wifi or wireless product - getting the queue theory right, and "out
there"as a basic component of everything, no matter what is layered on
top. This is *hard*. Work is just beginning.

Delivering a polished product with all the desirable features is
99.999% more work than that. Realizing, that in wifi, most of the
problems can't be solved at the AP, but in the clients and their
drivers, depressing.

If it were possible to assemble a funded team to produce an "openwifi"
product - or one as open as possible, as seems possible with the
mediatek chipset - I'd join it.

Same goes for producing open hardware ip for wifi to be, say,
co-joined with the risc-v or millcomputing work. But we're talking
millions of dollars here even with university help. I helped fund this
- http://www.meshsr.com/ - to no nearly no avail, but there is a lot
of work taking place on the zynq derived FPGAs that one day might see
the light as ASICs.
http://www.mathworks.com/hardware-support/zynq-sdr.html?requestedDomain=www.mathworks.com

There are niche markets opening, in other spectrum bands (802.11ad),
and I still retain hope that UWB could take shape somehow, maybe as an
in-house backhaul... It may well be that micro-cells from 5G will be
the wave of the future.