Femtocell reference design from picoChip heralds new phase in wireless buildout

As the pressure of growing data traffic bears down on cellular networks, service providers are turning to femtocells as an achievable near-term solution. This growing interest is speeding the maturation of the femtocell market and triggering a scramble among silicon providers.

The proximate issue is network capacity, a crisis triggered by the huge data bandwidth appetites of smartphones. Not only are these handsets rich in Web-gnawing applications, but, as Interdigital fellow Bob DiFazio observes, "the new mobile Internet traffic is not optimized for limited bandwidth. HTML is a hundreds of times greater load than bandwidth-conserving formats like Widsets or Bluepulse."

But the options for increasing air bandwidth are few, and scattered along an uncertain timeline. The obvious step—upgrading the network to a more advanced air interface—is not easy. Service providers in North America are still struggling with the move from their already-creaking 3G networks to wider HSPA deployment. But other alternatives, especially LTE, are looming, and a five-year-away replacement for LTE is already in the planning stages. These are not short-term fixes.

"There are providers moving into HSPA+ today. But LTE is further behind. A few will have it next year, but not early. No one is rushing," reports Rupert Baines, picoChip vice president of marketing. In part, the providers’ deliberate pace is financial necessity. Credit markets are still all but closed, and the providers’ all-you-can-eat service plans have limited the free cash flow to finance any buildout. And the air interfaces have reached diminishing returns. "LTE is approaching the Shannon limit," DiFazio says. "Further increases in bandwidth are going to have to come from more complex network management or from more cells."

That’s where femtocells come in. Originally, many saw the book-sized miniature base stations as a way to inexpensively fill in the coverage shadows in advanced services, whose multi-GHz frequencies are ineffective at penetrating obstacles such as walls or leafy trees. Demanding users may find their new smartphones are off the network altogether in their own homes. And the actual data bandwidth on services can drop by a factor of 100 at the edge of a cell, according to DiFazio, leading to an apparent loss of service for data-intensive applications. So sticking a femtocell keyed to the subscriber’s and his friends-and-family’s handsets in the subscriber’s home could significantly help with coverage complaints.

But now, service providers are starting to examine femtocells in a different light. If a user enters the network via a femtocell—which connects directly into the wireline Internet instead of relaying through the nearest base station—that user’s traffic doesn’t take up precious base-station capacity. Deploying femtocells in large numbers is almost as good for the service provider as increasing the capacity of his base stations. And there’s no new expensive upgrades for the provider to absorb: he can make the subscribers pay their own way and even install their own boxes.

This benefit naturally has caused a surge of interest in femtocells. But the boxes are not an easy design problem. A femtocell has nearly all the functionality of a full base station, just scaled down in number of channels and transmit power. And it must be cheap: Vodafone in the UK is asking about £120 for a box.

As you would expect, base-station silicon vendors are scaling down their architectures to reach the femtocell market, and both systems houses and SoC developers are working on dedicated silicon. But the market may already be evolving beyond the point where bare silicon is enough. This morning picoChip—which arguably had the first femtocell silicon solution in the open market—announced a full femtocell reference design based on its picoXcell SoC. The reference design is a fully ODM-ready kit including RF, baseband, control and networking silicon, board film, software—the works. Baines says that the femotcell market is rapidly moving from the leading wireless OEMs to the smaller, less technical houses that need a turnkey package.

Nor is the evolution stopping. Baines says that now interest is broadening beyond residential femtocells to what he calls Greater Femtocells—boxes with essentially the same functionality, but with extended channel capacity and range to handle moderate-sized areas with coverage problems, such as roadside rest areas, parking lots, and the like. "The use pattern would be similar to WiFi hubs," he explained. And beyond this, look for merged devices for enterprise use, offering femtocell services, integrated PBX functionality, and connection back into the enterprise’s IT organization. In this scenario the femtocell becomes not personal property, but a portion of the enterprise network. Such steps will test the scalability of any femtocell silicon.

In concept, the femtocell is a shrunken base station. But the technical and marketing differences are sufficient for the leading edge of femtocell deployment to open the door to all the players, not just established base station providers. It’s not at all clear who the winners will be, but this is clearly going to be a fast-moving market.