Will NanoAPs and HomePlug kill the Femtocell?

This is going to be the year of the Femtocell.  At least that’s the message that the industry is putting forward. Next month at the Mobile Congress in Barcelona, the industry is likely to be united in singing off that particular hymn sheet. However, an RFQ from a network operator that was put out just before Christmas suggests that opinion might not be as solid as the industry hype portrays. Rather than looking for femtocells, this particular operator was contemplating the deployment of small 802.11 access points around the home, connected together and to the broadband line using HomePlug. The implication is that instead of providing a personal 3G cell in the home to compensate for their lack of indoor coverage, they’d prefer to flood it with Wi-Fi. It’s an interesting approach…

The concept of femtocells is not new; it’s just the latest step along an unsteady route to convergence. The mobile industry has always struggled with the issue of getting rid of the fixed telephone line within the home, or of finding a way to own the calls made over it. Ten years ago, the first approach to convergence (which is the word which mobile operators euphemistically use when they mean stealing your landline), was made by introducing a dual mode GSM / DECT handset. That failed to catch on, partly due to the packaging constraints of getting everything into an acceptable size. Instead of refining the concept, the industry ditched it and moved to a concept called UMA (Unlimited Mobile Access) which entails putting a long range Bluetooth or Wi-Fi capability into the handset and using this to communicate to a wired access point in the home. UMA goes further than DECT, as it handles the process of taking incoming calls which can’t reach you over the mobile network and rerouting them through the landline, then via the short range wireless link to the handset. That’s the concept of hand-over, where the network recognises the fact that you are moving in or out of range of the local access point or cellular network and routes your calls accordingly. It means that you always appear to have a mobile connection. Once again, the early attempts at implementing the technology were not really up to the job, and with the exception of a few worthy, but largely ineffective efforts to sell these systems to the public, they have been consigned to the “bad ideas” scrap heap while the industry has once again moved on.

The transition to 3G has added urgency to finding a solution, as it has brought a new problem – the issue of poor in-home coverage. Many users, particularly in rural areas have discovered that their indoor coverage with 3G is considerably worse than it was with 2G, and in a significant number of cases non-existent. That’s anathema to the network operators, as a large percentage of calls and mobile downloads occur at home. The problem is that it’s not economically viable for them to add enough additional infrastructure to improve coverage. Hence the approach of putting a miniature base station into a subscribers home and using the user’s broadband link to connect it back to the network has been seen as an attractive option. That’s what a femtocell is. And as it appears on the operator’s network as a mobile cell, handover should be moderately easy to accomplish.

The question is “Who funds it?” Today the cost of the bill of materials for a femtocell is around $100. Network operators, who have failed to provide a strong enough signal for users to make calls (which means they lose potential income), have come up with the neat but non-intuitive idea of asking customers to pay for the femtocell themselves. Not surprisingly, this has had limited success. The customer argues that they signed up for a service that promised coverage, so why do they now need to pay more for the privilege of getting it to work at home? To address this, operators are hastily thinking up new uses for the femtocell to try and make them more “attractive”. Which means trying to find benefits for them that extend past fixing the issue of poor coverage.

Foremost amongst these additional services is the belief that consumers will want to use their phone not only for voice, but also to consume high speed data services on them using the local femtocell connection. There’s a couple of ulterior motives here. The first comes from the observation that shows that users are more likely to experiment with and use such services at home. If they work there, the strategists argue, customers are more likely to use them when they’re out and about, increasing their data spend. The second, and even more attractive motive, as far as the network is concerned, is that the data link to the femtocell is made using the customer’s own broadband line, so it costs the network operator nothing to deliver the data. So by making it easy to experiment at home, at no cost to the network, the network will reap the benefit of more data use when the user leaves home.

Of course, if this strategy works it implies a growing use of data services as users start to understand them. That threatens to overload the network infrastructure, making these services slow and less attractive just at the point they’re becoming popular, which can only be solved by more infrastructure investment. But nobody talks about that at this stage.

Another benefit in the operator’s eyes is the possibility of introducing family plans. This arises from the fact that each different operator would need their own femtocell in your house. So if all the members of the family had contracts with different operators, (which isn’t unusual), they’d need as many femtocells as they have operators. Operators would love to lock all the family to one contract and the femtocell is a perfect tool to offer attractive pricing plans to this end. But even then the operators are loath to pay for the hardware. They insist that they need a way to make sure the femtocell is paid for, so that they’re not held hostage to the fickleness of users who churn their network.

Which brings us back to the nanoAP and Homeplug. It’s an approach that steps back from the current femtocell fascination to reconsider Wi-Fi and UMA as the means to connect any number of handsets inside the home. It has drawbacks – it’s unlikely to be as easy to set up, particularly on the handset. In contrast, femtocells should just work. Wi-Fi won’t, although an intelligent operator could go a long way in terms of adapting the user interface to make it much easier. The key benefit is that this approach is much cheaper. It also extends the users own broadband network, which means the user is more likely to pick up all or part of the cost.

They’re cheaper because they’re simpler. Today, the raw cost of a GSM capability in a volume of low millions of units is about $11. Upgrading that to 3G and adding the additional transmit and receive circuitry to do a proper femtocell probably doubles that. Add the additional processing to control it and you’re up to $30-$40. But that’s in millions of units; in the early years costs are likely to be closer to $100.

In contrast the nanoAP uses everyday Wi-Fi and Homeplug silicon and probably comes in at a component cost of no more than $15 for each access point in quantities of tens of thousands. If the user configuration can be sorted and UMA made to work properly, that’s a very attractive alternative. From the user perspective it offers a similar capability in terms of connecting their mobiles; it also provides full Wi-Fi coverage throughput the home for their existing broadband connection and it can support handsets running on multiple networks. Plus it will go on being useful if they change network providers. Even if their new provider doesn’t offer UMA capabilities, they can still load a mobile Skype client onto their handset and get cheap outgoing calls. So they’re more likely to bear the price tag.

It will be interesting to see if it happens. A number of companies have staked their future on femtocells taking off and this is the year when they’re going to have to prove to investors that this strategy is valid. It’s not at all obvious that they saw Wi-Fi as a contender. Even if they did, they probably didn’t see multiple, cheap access points that just plug into mains sockets and connect using HomePlug connectivity. It’s a terribly attractive approach, albeit needing the backend UMA issues to be sorted out. If one operator is seriously considering it, then there may be more making the same calculations behind the scenes. If that’s the case, it could be a very rocky year for the femtocell story.