One of the big promises of the IP multimedia subsystem architecture has always been fixed/mobile convergence, or FMC, as vendors have sold carriers the idea of integrating their wireless and wireline networks into a single service. The first result of this integration was supposed to be the dual-mode phone, a Wi-Fi/cellular hybrid that would allow mobile calls to tunnel through the Internet, turning a LAN and a broadband connection into a node on the carrier network, the home phone into a mobile phone and vice versa. The promise of FMC hasn't changed, but its direction certainly has. The mobile industry is no longer talking about dual mode. Instead, by 2010, FMC could be much more about femtocells.

Instead of using another radio technology (Wi-Fi) to carry the voice call, femtocells act as mini cellular base stations that, like dual-mode technologies, use a broadband connection to backhaul the call over the public Internet. They both accomplish the same goal, but femtocells have some clear advantages.

Dual-mode solutions require dual-mode phones. And although the number of Wi-Fi-enabled cellular handsets is growing, femtocell networks would not require any handset replacement — the same radio that connects over a macro-cellular base station works to the femtocell. There is a trade-off between infrastructure and handsets, though. There may be no cost to exchange handsets, but the carriers have to provide the femtocell.

Companies such as PicoChip, which are creating the radio chips powering femtocells, are promising to deliver goods that will bring the total cost of billable materials to $100. If the industry can achieve such low costs for a home gateway, deploying them becomes simple math, said Jeff Brown, CEO of RadioFrame, which is adapting its RadioBlade architecture for femtocells. A carrier could spend $100 a head to replace the handsets for a family of four or shell out the cost of a femtocell, Brown said.

Beyond the economic equation, though, femtocells are appealing because they're familiar, Brown said. Instead of supporting a new network technology, carriers are merely extending their networks into the home, creating the tiniest of cells. They aren't dealing with different spectrum, different radios or different protocols, nor are they dealing with an open public network, Brown said.

“I always have worries about putting what amounts to a licensed-spectrum technology on unlicensed spectrum,” Brown said. “I've worked in metro areas where there is so much noise on the 2.4 GHz bands [Wi-Fi spectrum] because you can see thirty access points wherever you stand.”

Sprint is one company that explored a dual-mode solution and has now shifted its focus entirely to femtocells. According to Iyad Tarazi, vice president of network development for Sprint, it already has an IMS architecture deployed and is migrating all of its networks to a common IP core, making a femtocell architecture a perfect fit.

“We're looking much more aggressively at femtocells on both the CDMA space and WiMAX,” Tarazi said. “It's not a question anymore of how to do it. We already know how to do it. It's a question of when we roll it out.”

Sprint's IMS-based network gives it a flat IP architecture, which would allow it to scale a femtocell deployment to almost any size. Although femtocells can be deployed on a legacy network, they are still miniature base stations that require communication with a radio network or base station controller. Since femtocells are designed to provide coverage to a single household, there could be tens of thousands of them in any dense urban square mile, requiring the controllers handling mobility management and radio resource management to scale up with them — a problem that a flat IP network with intelligent femtocells communicating back to the IMS core could avoid.

Although scale is a definite positive for Sprint's network, it isn't the carrier's first consideration, Tarazi said. Sprint is looking first at extending network penetration, reaching homes or businesses that may get weak access to the network or areas at the edge of its coverage area. It remains to be seen, Tarazi said, whether femtocells could evolve into a personal data and voice node in every home, but there are several other business cases in which a femtocell would make sense, such as in hot spot deployments.

A femtocell acting as a hot spot in a coffee shop or airport lounge would not only have the benefit of adding increased cellular voice and data coverage in a highly trafficked area, but it wouldn't require customers to access a different network. There would be no need to fool with registration screens or separate data subscriptions at every hot spot, Tarazi said — if you're on the Sprint network, you're on the Sprint network.

Europe has embraced femtocells; the first commercial deployments of the technology are expected by the end of year there. O2 and Vodafone have expressed interest in femtocells as well as several other European carriers. However, many of those operators haven't yet made the investments in IMS Sprint has or are much further behind in their deployments. But that certainly doesn't make femtocells out of reach. The reason is a protocol they all became familiar with in the dual mode days: UMA.

Although unlicensed mobile access is often associated with the Wi-Fi/cellular phone, it's really a protocol that will work on any FMC network deployment. UMA essentially tunnels the cellular call over the public IP network to a gateway where the signal joins the circuit-switched network as a normal call. There's no reason why that same technology can't be applied to femtocells, said Steve Shaw, director of marketing for Kineto Wireless, a UMA software and controller vendor. The difference is, Shaw said, the software resides in the femtocell, not in the handset itself.

Shaw said that rather than competing, UMA has been one of the drivers pushing femtocells forward. In principal, the femtocell made sense, Shaw said, but there wasn't a standardized protocol to connect it back to the network. While the industry is chattering about IMS and connecting to the femtocell with session initiation protocol (SIP), there have been no overarching industry guidelines for such an architecture.

“We pitched that side of the story to carriers that said they had no interest in dual mode, and they were receptive,” Shaw said. “There is a lot of talk about SIP and IMS, but using SIP to connect to femtocell is completely undefined. Just saying SIP doesn't mean a solution will appear.”

And although the flat IP networks and SIP-based communications will eventually arrive, carriers are still far off, Shaw said. Meanwhile, femtocells are ready now, and UMA gives them a way to communicate directly with the R4 switching infrastructure they have in place, Shaw said.

Not everyone is hot on the idea of UMA. Tatara Systems is promoting an all-IP approach to femtocells that take advantages of what IP and IMS elements carriers already have in their networks, regardless of whether they may be far from constructing a full next-generation network. Jonathan Morgan, vice president of marketing for Tatara, said that an IP or SIP-based call on the femtocell may need to be converted when it hits the legacy network, but he pointed out that UMA technically converts the GSM call into voice over IP (VoIP) and has to go through a gateway like an all-IP solution.

In addition, an IP solution with the intelligence concentrated in the femtocell brings a wealth of other opportunities to carriers. As carriers deploy more of those elements, enhanced features and integrated services can be extended to the femtocell easily, Moran said, just as VoIP providers like Vonage can add IP applications to their basic VoIP services. In addition, that convergence functionality can be extended all the way to the femtocell, where broadband, home phone and even video services could be integrated and controlled, making the femtocell more of a universal home gateway, rather than just a mini-base station.

“With an all-IP approach, you can basically combine all of the services in the home,” Morgan said. “From a mobile provider's point of view, you can basically own the home.”