The recent spate of news on fixed-mobile convergence (FMC) deployments (mostly trials) has caused me to ponder the differing technology approaches being followed. Having injured my back in a freak airport incident earlier this week, I had some time to survey the landscape. Here is a brief compendium of my research:

• Both Sprint spin-off Embarq and T-Mobile USA have launched UMA (unlicensed mobile access) systems in recent months. The Embarq press release is here. Details of the T-Mobile trial are here, and an annoying flash presentation is here. Both systems use 802.11 WiFi access points to allow users to connect from home or the office to make calls at landline rates. With T-Mobile, service extends to any T-Mobile public hotspot (such as the local Starbucks).
• BT has been running an FMC service called Fusion since late-2005. Fusion uses bluetooth rather than 802.11 for in-home and office access. Calls made when connected to the bluetooth access point are billed at BT’s landline rate. An early review of the service is here. BT claims that a WiFi version of the service is in the offing.
• UK provider PMN just launched a low-power GSM network that lets users connect to in-building GSM picocells with existing GSM phones and a PMN-supplied SIM card. The service runs on licensed spectrum designated for low-power gear. There are thirteen other UK companies that won similar low power GSM licenses and many other picocell trials have been announced.
• Not so fixed picocell deployments have been announced by OnAir (a partnership between SITA and Airbus) to allow passengers to make calls while in-flight. The British airline RyanAir is among the first to jump on the bandwagon, as announced here. (BTW, check out the top level domain in www.onair.aero – .aero. I’ve never seen that in use before!)
• Although no commercial trial is yet announced, a home version of the picocell, called a femptocell, has been produced by chip-maker picoChip. Companies such as Ubiquisys and ip.access build the gear and are working with service providers to put femptocells into commercial use.

The unlicensed approach (referred to as UMA or GAN by the 3GPP) is obviously beneficial for operators that do not have spectrum rights in a given market. It is also a potentially effective way to backfill a network where licensed coverage is spotty. However it requires the use of unproven dual-mode handsets to connect to WiFi or bluetooth access points. If you have ever had trouble connecting your PC to a WiFi or (worse) bluetooth access point, you will recognize this as a potential customer service nightmare.

The picocell/femptocell approach, on the other hand, works with existing GSM handsets and uses standard SIP to connect upstream. But it requires that specialized mini GSM base stations be deployed in homes or offices. This is logistically challenging from a provider planning standpoint – especially since a single household or office might have multiple base stations if its occupants subscribe to different providers.

It is still unclear to me what the regulatory treatment might be outside of countries like Britain which have carved out specific licenses for this type of low power, private GSM service. Would T-Mobile’s existing US spectrum licenses allow it to use pico/femptocells instead of UMA?

Props to Dean Bubley, who appears to be the de facto expert in all matters FMC. His Disruptive Wireless blog is a must-read for anyone wishing to follow this part of the industry.