Femtocells, the indoor cellular base stations that are tipped to be a big seller in the next few years, could shake up the prospects of all wide-area network technologies, according to last month's Femtocells Europe 2008 conference in London. But before they appear on today's 3G networks, they have issues to overcome.

No trials yet

A femtocell, as Techworld readers know well by now, is a low-power indoor 3G base station, that uses the subscriber's own broadband to increase coverage in the home or office. It connects voice calls back to the mobile provider's network, and data services to that network or directly to the Internet.

There are still no public trials of femtocells and, despite high hopes a year ago, none were announced at the Femtocells Europe conference. Instead, French operator SFR announced delays to its femto plans. and others dealt at length with the difficulties.

Why bother with femtocells at all? They promise great benefits. Operators could give or sell 3G users a femtocell to provide faster data indoors, longer battery life (because close up to a base station, the handset will need use less power) and new services such as in-building media. At the same time, other users outside the house would benefit, because the femto offloads traffic, freeing up the macro net.

If the benefits are that obvious, why are we still seeing no visible activity? In 2007, ABI Research predicted 36 million femtocells would be shipped (and shared by up to 152 million users) by 2012. Analyst Stuart Carlaw is predicting a later growth, but still expects 45 million units shipped in 2013. "Femtocells will take a little bit longer to come to market, with 2010 being the year of plenty."

Standards delays

Part of the trouble is the standards. "T-Mobile has concerns about progress of femto standards," Zhongrong Liu, head of RN (radio network) technology strategy at T-Mobile told the conference. "Originally 3GPP standards weren't written with millions of tiny cells in mind."

Despite the Femto Forum's progress on standards, anyone doing things with femtos right now would have to take a non-standard approach. Those involved with femto standards make an analogy with the early days of DSL, fixed line operators bought into non-standard equipment and costs were high.

In Liu's view, the inclusion of femto standards in the 3GPP's release 8 specification due at the end of 2008, is only the start, and the fully complete optimised femto standard won't exist till the end of 2009. In particular, operators need remote management to avoid costly customer service issues.

Cost - the $100 target

The other often quoted issue is cost. Operators are demanding femtos that cost less than $100, and hardware vendors promise they can do it. But can they? If one factor changes, such as the royalties charged for proprietary technology, it could alter the picture, Carlaw warned.

If femtos cost $128, and each femto generates a $20 monthly subscription, they become profitable in 2011, he predicted, but high technology licences will add to the cost and delay profitability. "The point of profitability may be pushed further out by IPR issues," said Carlaw. "High cumulative royalty rates can drive innovation out of the market," he continued, citing the market for CDMA based mobile phones and warning that an industry which treats femtos like handsets could spoil everything.

Another issue is how femtos integrate with other equipment and services. Firstly, they have to work well with the macro network. "Femtocells can create holes in the macro coverage," warned Rasmus Hellberg, director of technical marketing at Qualcomm. Users passing a house with a femtocell might find it interferes with their signal from the macro network, he warned.

The answer, broadly, is intelligence in the femto, including power management to limit the signal to inside the house, and the ability to move to a different carrier frequency to avoid local macro cells.

That must add to the cost of the devices.

Integration with the home

Integration with the user's home equipment is another issue. Carlaw's predictions suggest that femtos will be gradually integrated into other equipment, becoming part of the home broadband gateway, but most users don't want to hook up another box with Ethernet. It's also questionable whether femtos can assume free use of broadband which comes from a different - and possibly rival - provider, who may decide to throttle the unexpected traffic.

"The early implementations have to be integrated," said Paul Callahan, vice president of business development at femto maker Airvana. "They are also going to start with operators who own the wireline." At the very least, that limits the operators who are likely to jump in quickly.

Meanwhile, integrated devices may take longer to produce, and will affect the cost. Most standalone femtos use specialist silicon from picoChip, but Thomson - which makes the BT HomeHub and other gateways - adopted Airvana's technology for a femto sub-module for its gateways, which it hopes to use in operator trials when they finally start.

And femtos may also not work all that well with existing handsets. Ironically, by making sure that they are compatible with today's handsets, vendors have had to make femtos that (to those handsets) are indistinguishable from the macro network, and that limits the possibilities of femto-specific applicatinos.

Dean Bubley has warned that vendors should sort the issue out now, but they are relaxed about it: "Today's handsets aren't optimised for femtocells, but they still perform vastly better on those femtocells than they do on the macrocells," points out Rupert Baines, vice president of marketing at picoChip.

3G femtos - when or if?

Femtocells still have a sense of inevitability, and the potential to add lots of value to 3G networks, but given all these factors, there are uncertainties.

Another huge issue, is the fact that, while all this is going on, the mobile industry is bracing itself for a major technology change, from 3G networks, to all-digital 4G networks with faster access speeds.

Delays to femtocells will pitch them into that already tangled web. If 4G femtos are the real issue, what will that mean? That's an issue to which we will return.