5 ms·

OpenBTS: Cheap, open source cellular network

121 points by nir 17y ago

vaporstun 17y ago
Looks very cool, excited about the possibilities, but my first thought was "FCC violation." Did a quick Google search, looks like they did a test at Burning Man in 2008 and got local FCC approvals: http://lwn.net/Articles/297038/ http://lwn.net/Articles/297038/
- randallsquared 17y ago
  This seems very similar to the magicJack femtocell except for not having to use them, so if magicJack can get FCC licensing for their any-GSM MitM device, then surely this could also.
- nimrody 17y ago
  One more problem: Patents. GSM implementations require licensing IP from several major manufacturers (Nokia, Ericsson and others). As long as the project is small they probably won't bother. But just try basing a femtocell product on OpenBTS and they will be knocking on your door soon.
  - nir 17y ago
    I suppose something like this would find much of its use in places where neither FCC nor patents have a lot of clout.
    - borism 17y ago
      yeah, like here in Europe :)
    - vdm 17y ago
      If you dig deeper on OpenBTS, the primary motivation of the project is altruistic, rolling basic coverage out in the developing world where there is no current economic incentive, or respect for IP. That's not to say that it would be disruptive in developed markets as well, or course.
cnvogel 17y ago
There are currently (to my knowledge) two working open-source/free-software implementations of the GSM stack. OpenBTS (http://openbts.sourceforge.net/ http://openbts.sourceforge.net/) puts everything, beginning at the modulation/demodulation of the radio-frequency, into your PC and uses a USRP (http://www.ettus.com/products http://www.ettus.com/products) as the radio-frequency frontend (~$2000). OpenBSC (http://openbsc.gnumonks.org http://openbsc.gnumonks.org) uses a different approach and starts implementing at the Abis-interface, which is what the Base Transceiver Station (the thing with the antennae...) connects to. Here you don't need the USRP but a surplus GSM BTS (currently around or a little less than $500/€500 and the necessary E1-interface (for the Abis-link) is $200/€200, afaik). OpenBSC has been successfully tested at https://wiki.har2009.org/page/GSM https://wiki.har2009.org/page/GSM and http://events.ccc.de/congress/2009/wiki/GSM http://events.ccc.de/congress/2009/wiki/GSM using officially allocated frequencies.
toisanji 17y ago
This would be awesome to try and build my own magicjack femtocell. I wonder what hardware is needed and how much it would cost to do.
ashish01 17y ago
I have been interested in software radio for some time now but lack the technical knowledge to understand it fully. Can someone point me to links that can help me understand this ? I am sure it will be helpful to many people. From what I have read one needs a Universal Software Radio Peripheral (USRP) for sending and receiving radio signals but searching for this leads to things like daughter boards etc. which I did not understand. Can someone explain it here ? Thanks
- borism 17y ago
  Well, it's a long story, but I think in essence software radio is to communications is what integrated circuit is to computing.
- unwind 17y ago
  The wikipedia page on the USRP seems to explain it pretty well: <http://en.wikipedia.org/wiki/Universal_Software_Radio_Peripheral> http://en.wikipedia.org/wiki/Universal_Software_Radio_Periph.... It also goes into the issue of daughterboards, although not in-depth. The reason for daughterboards is given as "[...] designed to be flexible, allowing developers to make their own daughterboards for specific needs with regard to connectors, different frequency bands, etc."
jonnycowboy 17y ago
I wish they made a (much) cheaper USRP specifically for GSM, I wouldn't use the rest of the radio functionality (AM, FM, etc).
- cnvogel 17y ago
  The USRP doesn't do any AM, FM,... It's only a very fast ADC/DAC. Some of the additional modules have up-/downmix capability, though.
jrockway 17y ago
My dream of creating a GSM network called "Free Public Wi-Fi" is one step closer to fruition...
- blasdel 17y ago
  Almost all of those non-AP networks you see (usually with that name, and never with a route to the internet or dhcp) are due to a pernicious Thyphoid Mary behavior in Windows' wireless networking UI. In at least Windows XP, if you ever join a "computer-to-computer" network, your computer will advertise that network (or the last such network) whenever it is not connected to another SSID, in perpetuity. Airports are the primary breeding ground for this disease.
  - jrockway 17y ago
    Yeah. And I think it would be amusing to see your cell-phone connected to that network instead of "AT&T" or "T-Mobile" or whatever :)

121 points by nir 17y ago

Looks very cool, excited about the possibilities, but my first thought was "FCC violation." Did a quick Google search, looks like they did a test at Burning Man in 2008 and got local FCC approvals: http://lwn.net/Articles/297038/ http://lwn.net/Articles/297038/

randallsquared 17y ago
This seems very similar to the magicJack femtocell except for not having to use them, so if magicJack can get FCC licensing for their any-GSM MitM device, then surely this could also.
nimrody 17y ago
One more problem: Patents. GSM implementations require licensing IP from several major manufacturers (Nokia, Ericsson and others). As long as the project is small they probably won't bother. But just try basing a femtocell product on OpenBTS and they will be knocking on your door soon.
- nir 17y ago
  I suppose something like this would find much of its use in places where neither FCC nor patents have a lot of clout.
  - borism 17y ago
    yeah, like here in Europe :)
  - vdm 17y ago
    If you dig deeper on OpenBTS, the primary motivation of the project is altruistic, rolling basic coverage out in the developing world where there is no current economic incentive, or respect for IP. That's not to say that it would be disruptive in developed markets as well, or course.

cnvogel 17y ago

There are currently (to my knowledge) two working open-source/free-software implementations of the GSM stack. OpenBTS (http://openbts.sourceforge.net/ http://openbts.sourceforge.net/) puts everything, beginning at the modulation/demodulation of the radio-frequency, into your PC and uses a USRP (http://www.ettus.com/products http://www.ettus.com/products) as the radio-frequency frontend (~$2000). OpenBSC (http://openbsc.gnumonks.org http://openbsc.gnumonks.org) uses a different approach and starts implementing at the Abis-interface, which is what the Base Transceiver Station (the thing with the antennae...) connects to. Here you don't need the USRP but a surplus GSM BTS (currently around or a little less than $500/€500 and the necessary E1-interface (for the Abis-link) is $200/€200, afaik). OpenBSC has been successfully tested at https://wiki.har2009.org/page/GSM https://wiki.har2009.org/page/GSM and http://events.ccc.de/congress/2009/wiki/GSM http://events.ccc.de/congress/2009/wiki/GSM using officially allocated frequencies.

toisanji 17y ago

This would be awesome to try and build my own magicjack femtocell. I wonder what hardware is needed and how much it would cost to do.

ashish01 17y ago

I have been interested in software radio for some time now but lack the technical knowledge to understand it fully. Can someone point me to links that can help me understand this ? I am sure it will be helpful to many people. From what I have read one needs a Universal Software Radio Peripheral (USRP) for sending and receiving radio signals but searching for this leads to things like daughter boards etc. which I did not understand. Can someone explain it here ? Thanks

borism 17y ago
Well, it's a long story, but I think in essence software radio is to communications is what integrated circuit is to computing.
unwind 17y ago
The wikipedia page on the USRP seems to explain it pretty well: <http://en.wikipedia.org/wiki/Universal_Software_Radio_Peripheral> http://en.wikipedia.org/wiki/Universal_Software_Radio_Periph.... It also goes into the issue of daughterboards, although not in-depth. The reason for daughterboards is given as "[...] designed to be flexible, allowing developers to make their own daughterboards for specific needs with regard to connectors, different frequency bands, etc."

jonnycowboy 17y ago

I wish they made a (much) cheaper USRP specifically for GSM, I wouldn't use the rest of the radio functionality (AM, FM, etc).

cnvogel 17y ago
The USRP doesn't do any AM, FM,... It's only a very fast ADC/DAC. Some of the additional modules have up-/downmix capability, though.

jrockway 17y ago

My dream of creating a GSM network called "Free Public Wi-Fi" is one step closer to fruition...

blasdel 17y ago
Almost all of those non-AP networks you see (usually with that name, and never with a route to the internet or dhcp) are due to a pernicious Thyphoid Mary behavior in Windows' wireless networking UI. In at least Windows XP, if you ever join a "computer-to-computer" network, your computer will advertise that network (or the last such network) whenever it is not connected to another SSID, in perpetuity. Airports are the primary breeding ground for this disease.
- jrockway 17y ago
  Yeah. And I think it would be amusing to see your cell-phone connected to that network instead of "AT&T" or "T-Mobile" or whatever :)