The Unlicensed Mobile Access (UMA) standard, looks less and less exciting on phones. It could appear on femtocells - but is this a new territory to conquer, or a cosy retirement home for a technology that never really flew?

UMA was first proposed as a means to allow dual-mode phones to connect to the cellular network over the Internet, using indoor networks such as WiFi to extend mobile coverage indoors and offload traffic from operators' crowded base stations.

The technology was criticised for offering a limited version of VoIP, controlled by the cellular operators, compared with IMS and SIP. For a long while, the only version actually available was BT's Fusion, which uses Bluetooth for the indoor portion. Now, BT has finally announced the corporate WiFi version, and Orange France has just announced a converged service called Unik, also based on WiFi.

Around three other operators are launching WiFi versions this year, according to predictions by Kineto, the software company that is the main UMA flag-waver: TeliaSonera in Scandinavia, T-Mobile in the US and Telecom Italia in Italy. Kineto maintains that all is well, but analysts tend to be sceptical, as other ways to run VoIP services are emerging, while these services trickle out.

If a user has a WiFi handset, it can run a SIP stack, and support anyone's VoIP service. It could run several. Why should a user buy their VoIP from their handset provider, in the way UMA envisages, when there's plenty more around? To take one example, Truphone, which was in May, offers a VoIP client for the Nokia E61, which sits alongside the GSM/3G voice. The application has moved to a full beta stage, and currently offers subscribers free calls to any landline in forty countries - far better than the offer from NT's Fusion and the rest.

Converged handsets with UMA can be subsidised to make them more attractive, but like plain BT landlines, they'll only be used by people who can't be bothered to find a better deal. The number that get UMA handsets will depend not on how good the UMA service is, but how hard it is to move to a better deal. A first look at the Truphone service suggests it will ultimately be very easy.

A new frontier? At this point, a new approach seems to be appearing, which might help out UMA. Low-power indoor cellular access points are emerging to improve indoor coverage; like converged WiFi/cell arrangements, they use an indoor network to get signals where 3G doesn’t go, and offload traffic from the carriers wireless network using the subscriber's own broadband for backhaul, but the indoor wireless technology is the same as the other one - this access point is just a smaller version of the base stations outside on the cellular network.

The providers, like Ubiquisys, call them femtocells, trying to go one further than the pico-cells which are proposed for the DECT guard band in the UK.

Unlike WiFi schemes, these use ordinary phones, and the base stations are in the licensed spectrum, so they have to come from the operator. The femtocell has to connect back across the Internet to the operator's network and there are three options for this: SIP, a 3GPP protocol called Iub, and UMA. Unsurprisingly, Kineto thinks UMA is best.

"Today's mobile networks were not designed to handle hundreds of thousands of APs on the network," says Steve Shaw, director of marketing at Kineto. "UMA minimises network disruption and scales tremendously," he says. Base station controllers expect a few base stations with a lot of traffic on each; UMA equipment attaches to that, and is designed to handle thousands of base stations, in homes and offices, which are each used very lightly.

Another benefit is that UMA includes built-in encryption, over an IPsec tunnel, he says. Ubiquisys' base stations can include a SIM card, and appear to the network as a client device, acting as a proxy to the cellphones attached to them. This allows encryption to be applied easily, and means that conversion efforts are handled in the base station. UMA defines use of an IPsec tunnel, and EAP-SIM is used for devices, natural fit for UMA security process.

By comparison, he says Iub is implemented, but in non-standard ways ("it's intended for a small market of really large controllers"), and with no defined security, while SIP doesn't have anything defined for backhauling cells ("but obviously, it will solve world hunger").

UMA in the background UMA critic, Dean Bubley of Disruptive Analysis finds something good to say about UMA on femtocells: "These devices don't have user interfaces and lots of applications that need to be made "dual-mode" friendly," he points out, so integrating a UMA implementation in a femtocell would be easier than in a handset. The handset can continue unchanged.

But there's a sting here. This really is UMA looking for a new niche to survive in. "None of these non-phone UMA widgets actually move in and out of cellular coverage, which underlies the irrelevance of the much-vaunted cellular-WiFi handoff bit of the technology," says Bubley.

A UMA (unlicensed mobile access) client on a 3G femtocell might be UMA, but it wouldn't be mobile and wouldn't be using unlicensed wireless.