I came across the OpenBTS project which is building a GSM basestation using Open Source software and hardware. Call processing is handled using the popular Open Source Asterisk PBX, with lower protocol layers written from the 3GPP specs. Open source hardware from the USRP project provides a software programmable radio solution – the hardware kit costs $700. Prototypes have been tested (with appropriate spectrum test licences from the FCC) and calls have been made. There’s no obvious technical reason why this project couldn’t be embedded into a PC or linux server.
Companies like Digium (who sponsor Asterisk) and Dialogic (now part of Intel) have been selling PC-cards which provide telephony interfaces for years. These allow OEMs to build office PBX switches and other fixed line and VoIP telephony applications from standard PC server hardware. Potentially, this project could enable vendors to offer PC cards with full GSM basestation capabilities.
So could anyone then just build their own femtocell, by plugging a card into their home PC?
Issues that quickly come to mind are legal rather than technical. Operating any radio equipment requires a licence (mobile networks have paid billions for theirs), partly to raise money for the government but largely to avoid interference between different systems. Who would want to be responsible for police or emergency services radio systems failing when they are most needed?
Mobile telecommunications is heavily populated with strong patent portfolios, for which all equipment manufacturers have to pay licence fees. Many of the essential GSM patents are now getting quite old and due to expire – one of the reasons for the project to concentrate on this technology rather than 3G. Maybe some enterprising PC card vendor can negotiate an inclusive patent licence for their card which addresses this issue.
A potential option is for a low power spectrum licence to be issued either free (like DECT cordless phones at 1800Mhz) or with a licence fee per unit, perhaps also tied to the hardware card.
The project will also raise some security concerns. We’ve often seen WiFi enables PC’s pretending to be free WiFi services. This can even be done by viruses running on WiFi laptops without the owner being aware of it. Spoof GSM basestations could connect your call traffic and listen in to your calls. However, since the GSM SIM card security scheme does not pass shared secrets through the network, it is unlikely to crack or clone your SIM card and steal your calls.
I’ve spoken with an Asterisk expert who had been investigating PBX functionality sitting alongside a femtocell, so that standard in-house telephony functions might be provided this way. This is quite different from the approach above, and Asterisk would not have any GSM or mobility protocols. There have been suggestions that some femtocell vendors have prototyped this functionality within their products already.
The background information on OpenBTS indicates potential applications in remote areas in developing countries, such as solar powered sites which could provide wide area coverage for small communities at minimal cost. It may also find applications in developed countries. With heavy competition forcing cellsite vendors to steadily reduce their hardware equipment prices, (one reason why RAN vendor profits have been low or negative in recently), it seem less likely that this project will threaten the mainstream industry.
With the market focus being on 3G femtocells, which will provide high speed data as well as voice, commercial products are becoming available that are low cost, very easy to install and fully integrated with the existing mobile networks. An open source 2G femtocell is therefore more likely to appeal to techies and specialist/non-mainstream applications in the first instance.
This is one project to watch and see how it evolves.