Peer-to-peer vs Femtocells. Let the topology war begin.

Two of the new wireless technologies that have come to the fore this year are high speed peer to peer connectivity and femtocells.  Although they may not appear to have any obvious connection, I would argue that they do.  Moreover that connection is so strong that they will end up fighting a technology war between themselves for a key customer application.  That’s because they both have a major impact on the way that users transfer data between their personal devices.  Today these two technologies are barely aware of each other – they’re both too busy gazing at their respective technical navels and ignoring the user requirements.  Within twelve months, when they understand the real use cases they’re enabling, they may well be at each other’s throats.

They’ve grown from totally different requirements.  Femtocells solve the problem of mobile coverage within the home.  That’s an important need for 3G and an essential one for 4G (LTE) networks.  They’re needed, as the higher frequencies used by these technologies are poorer than 2G networks at penetrating the walls of our offices and homes.  Operators naturally don’t want to lose customer traction by having them put their mobiles down and reverting to a landline or the dreaded Internet VoIP – they want them to stay on network.  So they’re reluctantly accepting the need to deploy femtocells to provide the coverage that users have come to expect.

Femtocells work by providing a personal base station within the home.  However, rather than using wireless or a backhaul provided by the network operator they use the home owner’s broadband link.  When you get home, your phone detects the femtocell and switches from its previous base station to your personal femtocell.  The clever bit of the technology is in ensuring that this is a smooth handover and that the connection is limited to your registered handsets.  It lets your handset obtain broadband data rates that are at the top of the range for the mobile technology.  It also moves the phone off the operator’s wireless network.  That’s important as devices in a poor reception area can result in the overall speed of an LTE network plummeting.  For more details on that effect, which is not widely acknowledged, but a potential killer of mobile broadband services, read the excellent article by Moray Rumney. 

The problem with femtocells is in trying to work out who pays for them.  They currently cost around $200, with hopes that that will fall to below $100.  Network operators would like customers to pay.  Customers feel that’s an imposition, as they’ve already paid for their phone contract and consider that it’s the network’s fault if they don’t get the level of coverage they got with their previous handset.  It’s unclear where this particular battle line will end up, but it is resulting in a degree of innovative billing schemes and additional services being dreamt up by the femtocell community and mobile operators to try and sugar the inevitable pill.

Many of these extra services are based around “family plans”, where the femtocell is used by all members of the family and possibly by designated friends.  As there’s a pretty good guarantee of full data rates in the home, this can be tied to some attractive high volume content downloads, as these won’t touch the mobile network.  From a user viewpoint these will probably be much easier to use than trying to use Wi-Fi, or a cable and PC to transfer them to the handset.  That’s because everything works from the default phone interface.  Hence the belief that it will be attractive to the user as it’s is easier to use. 

In a parallel universe, the Wi-Fi, or Bluetooth and 802.11 communities have other ideas.  They’re trying to promote high speed peer-to-peer connectivity between devices to address the issue of moving content between devices.  The biggest announcement was from the Bluetooth SIG, whose recently released version 3.0 of their standard combines Bluetooth with 802.11 to increases the data rate between two devices to around 20Mbps.  UWB was also in the frame for this standard, promising to trump this by a factor of ten.  Unfortunately it’s been hit by political manoeuvring and the recession, with most of the supporting companies disappearing from the frame.   It’s part of UWB’s turbulent history, where it has repeatedly failed to get its timing right.

UWB aside, the promise of these technologies is that we’ll be able to transfer content quickly between devices.  However, there’s some concern as to whether anyone’s interested.  Five years ago, Bluetooth 1.2 could transfer data between devices at a speed of around 700kbps.  Nobody found that particularly exciting and picture transfer didn’t take off.  Today with Bluetooth 3.0 we can do it at thirty times that speed.  Whilst the chips companies have got excited by these new go-faster stripes, they’ve failed to notice that the camera resolutions have increased at almost exactly the same rate.  So it takes just as long to transfer the photo today as it did five years ago.  That makes it difficult to understand why it should be any more compelling an application now than it was (or rather wasn’t) five years ago.  Which brings you to the conclusion that high speed peer to peer, particularly if it involves learning a new interface, may be still-born.  The step change in speed and lower power consumption that UWB might have offered could have changed the equation, but that opportunity may now be lost.

Even if we make the dubious assumption that the data rates are fast enough, the other question that needs to be asked is whether this sharing model has come too late.  Or to put it another way, is ad-hoc transfer as dead as the dinosaurs?  We’re so used to using networks to send content to other devices, that it may be too late to change.  Most people will send a file to a colleague or family member as an email attachment, even if they’re in the same room.  The concept of sending it directly between two devices rather than via a network is surprisingly non-obvious to most people.  It died with floppy drives.  If femtocells make the network model easy enough, then peer-to-peer has the problem of breaking down an “I’ve always done it this way” barrier, regardless of whether it’s a better technical solution.

Today’s current fashionable applications only serve to reinforce that paradigm.  We publish rather than doing individual sharing.  Which is why users upload photos to web applications like Flickr and then tell their friends – it’s sharing for communities.  This model has plenty of room for networks to innovate, combining content movement with social networking alerts, which will further reinforcing this as the way users think.  By enabling this at home, femtocells once again plays to the network game. 

So although the topologies of femtocell and peer-to-peer appear diametrically opposed, in many ways both address the same problem – how to get large volumes of content on and off our mobile phones.  Of course, how we get it on in the first place is a moot point.  Some, such as photos and videos may be generated by the devices.  Much more won’t be.

Today there is only one good way of moving tens of Megabytes of data between devices, which is a cable.  That largely confines data transfer to when we’re at a desk with a PC.  That may be valid for printing photographs or downloading content for email attachments, but it turns the mobile device into a tethered one. 

Femtocells and the network approach offer a unified user experience, giving phone users a single interface for their content.  It’s vital for both industries to realise how important that is as they approach this battlefield.  Familiarity is independent of technical superiority and will trump it.  If a femtocell makes something simple, users will ignore the alternative of Wi-Fi, which could remove the need for Wi-Fi in a handset.  And there’s nothing else to keep Wi-Fi if an alternative, compelling download content interface and delivery mechanism appears.

Wi-Fi proponents need to be very aware of the danger of letting phone users retain the same interface without having to worry about network cost.  Nor should they forget that mobile phones have always been embedded devices – the key services on them work out of the box.  Wi-Fi and the PC industry in general have always struggled with that concept. 

Femtocells may also find other weapons to fight with, if they evolve past being simple base stations.  One of which would be to provide local network storage, so that they can become the family’s local content store.  Or they could allow us to share content between devices in the same locality, or with specific friends over the network.  There are some interesting approaches that designers can make, such as allowing free data transfer between devices that are on femtocells, rather than the wireless network.  It feels a technology far more open for technical and service evolution that peer-to-peer.

Of course, users may not care for either.  But if we really want to turn the phone into a personal content device, the battle is on to find out which will provide the connection.