OpenBTS on Android

[Antoine van Gelder]

@chrispaget writes about running a GSM base station from an Android handset:

http://www.tombom.co.uk/blog/?p=144

Though I'm not sure I agree with him that it is a somewhat pointless hack.

:-)

- antoine

[Antoine van Gelder]

http://www.tombom.co.uk/blog/?p=144

:-)

- antoine

--
http://7degrees.co.za
"Libré software for human education"

[wayne]

Antoine van Gelder wrote:
> @chrispaget writes about running a GSM base station from an Android handset:
>
> http://www.tombom.co.uk/blog/?p=144
>
> Though I'm not sure I agree with him that it is a somewhat pointless hack.
>
> :-)
>
> - antoine
>
> --
> http://7degrees.co.za

> "Libr� software for human education"
>
>
>
Antoine

Maybe I've lost the plot here, but this still needs a USRP as far as I
can make out, So what are you gaining?

It's basically using the Droid as the PC.

Wayne A

[Paul Gardner-Stephen]

Hello Wayne,

A very low power, compact, robust, battery powered and portable PC.
These all make it incrementally easier to deploy. Of course, there
are other options out there for doing this, but using a comodity
device saves a lot of leg work.
Also, the droid phone also provides WiFi, so could cross over into a
mesh potato network. This is very much in line with what Serval
wishes to achieve.

Paul.

[David A. Burgess]

Paul, Wayne, Antoine -

One of the reasons that Chris was able to do this is was that Harvind
and I had put considerable effort into running OpenBTS on a Gumstix
board. That was about a year ago. We added a lot of configurability
to the radiomodem to allow it to be trimmed down for short-range
applications. We made a lot of computational optimizations in the
signal processing code. We also made a lot of changes to the
threading and memory footprints to allow the code to run on machines
with small caches. We did not say much about the Gumstix port
publicly, but did we put all of those changes into the 2.5 public
release. Our experience was that we could run 2 active timeslots
with no radio channel equalization on a 500 MHz Overo, which is
adequate for femtocell or SMS-only applications in small spaces. Of
course, running OpenBTS on a small CPU shifts the power/size focus to
the radio itself, which is still big and still drawing about 15 W.
We are working on new radio hardware to improve that.

For full-range, 8-slot systems, our own workhorse platform is still a
mini-ITX PC with a 1.6 GHz Atom and a flash drive. The installation
in Niue draws 60 W total, including a power amplifier and the 802.11a
backhaul. The backup power supply is 3 deep-cycle batteries from the
local marine supply shop, which should keep the unit running for 2-3
days the next time a typhoon knocks out the power. That's not a
pocket-sized device, but then it is designed to operate over a range
of several km.

-- David

> --
> You received this message because you are subscribed to the Google
> Groups "village-telco-dev" group.
> To post to this group, send email to village-telco-
> d...@googlegroups.com.
> To unsubscribe from this group, send email to village-telco-dev
> +unsub...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/
> group/village-telco-dev?hl=en.
>

David A. Burgess
Kestrel Signal Processing, Inc.

[Paul Gardner-Stephen]

Hi David,

I continue to be impressed and amazed at the great work that your team
has carried out with OpenBTS to date.
While the power budget may still be significant, 60W is still much
better than 6-16KW for a commercial BTS, and I feel well within the
scope of reasonability for disaster relief and developing/remote
applications.

Once we have our stage 1 complete (meshing Android phones with Mesh
Potatoes to prove the de-risking of a fully distributed mobile and
fixed telephone network), we will begin seeking funding for, among
other things, incorporating OpenBTS into this equation, so that we
will have a fully distributed model for telephony that can use
existing mobile telephones as well as Mesh Potatoes and specially
configured WiFi equipped phones.

I have begun to explore our options for gaining continuing access to
some GSM900 spectrum here to ease development, which will be a
fantastic victory if I can achieve it.

We are tremendously excited about the possibilities that the
combination of OpenBTS, Village Telco and Serval creates, and look
forward to continuing to work with you all as we strive towards a
vision of appropriate, affordable, and accessible telecomms for all.

Indeed it is only about 48 hours before we begin our local shake-down
day, and then on the following day fly to Arkaroola for a serious
field trial and technology demonstration.

While I still have plenty of small jobs to do between now and then
(hopefully including sleep!), I can say with great satisfaction that
Dany and I have successfully achieved all that we set out to achieve:

1. We have Android phones participating in a Village Telco mesh with
Mesh Potatoes.
2. We have mesh potatoes and Android phones both sporting our
Distributed Numbering Architecture (DNA) that allows clients to claim
or choose their own telephone number and dial one another without
requiring any centralised service.
3. While still experimental, we have the user interface sufficiently
usable as to be practical. Dany has put in some great work getting
things like caller-id working, so that the mesh is beautifully
intuitive to use with DNA numbering.

Many challenges remain, but we have reached a significant mile-stone,
and we should all be pleased with our various contributions that have
bought us to this point.

Anyway, I should crawl off to bed and get some sleep.

Paul.

> > For more options, visit this group athttp://groups.google.com/