Doug believes passionately that metro-femto is the only way to achieve widespread high-capacity public mobile data service, giving operators who have already deployed residential femtocells a natural advantage. He also shares his views on LTE including the market with most potential, the close alignment between TD/FDD modes and SON (Self Organising Networks).
Picochip’s New Offices
Situated in the heart of the ancient city of Bath, only a stone’s throw from their smaller and outgrown office, Picochip’s spacious new office accommodation reflects the growing success of this international business. Founded some 10 years ago, it now has (research and development) offices in Bath, Cambridge (UK) and Beijing.
I have to say that it’s either ironic (or just reinforces the strong market need for their product) that the mobile phone coverage inside is so poor (unless you happen to be able to use their femtocell).
I spoke with Doug Pulley, their co-founder and CTO, asking specifically about their latest advance into public outdoor 3G femtocells (so-called metro-femto) and the early stages of LTE femtocells.
3G METRO FEMTO
The specifications remain quite demanding
LABS (Local Area BaseStations) are a class of mobile phone basestation which sit between the very large and long range towers (macrocells) and the small indoor residential and office femtocells. Often these are distributed at intervals of a few hundred metres apart in very high traffic areas.
Picochip have worked hard to engineer a solution that meets the LABS specification, announcing a fully compliant product in September this year.
Doug’s view is that the LABS specification wasn’t dumbed down enough from macrocells if it was really intended for 50m cell range. For example, it still includes the need to handle users travelling at up to 120km/hour and significant multipath effects (echos from surrounding buildings) which requires heavy baseband signal processing to cope with these and the resulting Doppler effects. Picochip have deliberately extended the capability to some 2-3km range to take advantage of this stringent standard.
As the range of a basestation increases, the window of correlation for signals from users at the cell edge also grows. This defines the range and hence the cell radius. The larger the window of correlation, the greater the processing power required to deal with it.
Macrocells will always be required, but femtocells will deliver the capacity
Networks will always require macrocellular towers for rural and sparsely populated areas, motorways, trains etc. but those are in place already.
Dense urban environments are where the battleground for good coverage and capacity now lies, especially for data. A large number of small, self-configuring and self-optimising basestations is really the only way to meet customer expectations.
Although the major basestation vendors are today offering modern, high capacity multi-RAT (2G/3G and 4G basestations in the same box) products, these are not targeted and not well suited to delivering the much greater traffic capacity required. There may well be a good business case to install these new products, which are much more energy efficiency and adaptable than previous and near end-of-life equipment. But there remains also a very good business case to deploy metro-femto for capacity.
Those operators who have already taken the step of residential and/or enterprise femtocells are ahead of their competitors. They can relatively easily adopt metro-femto, reusing the femto gateways they now have in place.
LTE LAB DEMO
Doug showed me their latest lab demo (pictured right). The equipment in the rack emulates a complete LTE end-to-end network plus extensive testing and monitoring gear. The LTE femtocell (one of the smaller boxes near the top) is connected to a commercially available LTE data dongle and standard Windows PC, proving that their technology is fully compatible with real, live products from other vendors.
Operating at 750MHz frequency with 10MHz bandwidth in FDD mode into a Category 1 commercially available data dongle, it supports the specified 10MBit/s downlink and 5Mbit/s uplink (but has gone faster).
We saw high quality video being sent and received bi-directionally without jitter or delays.
Would it work outside the lab environment I asked? Extensive conformance testing suggests that it would – some 25,000 automated tests have been performed to ensure it matches standards and meets regulatory and performance requirements.
We’ve seen the pace of commercial LTE launches ramp up in recent weeks: Verizon Wireless, Vodafone Germany and Telesonera to name but a few.
Doug believes that any of these could be overshadowed by China, which could quickly become a very large market. Perhaps 3G hasn’t been quite as successful there and China Mobile has some 100MHz of TDD spectrum at 2.3GHz which could be used for either 3G TD-SCDMA or TD-LTE. (This is in addition to the 2.5/2.6GHz band being adopted in many other countries, but not yet released for use in China.)
Won’t the added complexity of TD and FDD modes cause issues?
Doug believes that most vendors anticipate demand for both TD or FDD modes. It’s already embedded into silicon for handset devices and Picochip silicon supports either mode. There is very strong alignment in the specifications for TD/FDD at L1/L2 and L3 – they are structurally very similar. It is quite feasible that you could be using FDD mode one minute and handover to TD mode in a different environment the next.
What format will LTE femtocells most likely take?
Dual mode 3G/4G metro-femto would be the most effective way to introduce extra capacity in urban environments. It makes best use of all available spectrum, device compatibility and spectrum efficiency. Picochip’s earlier prediction of a sub-$70 BOM for such a device some years ago is now very much a reality. You can expect to see some dual mode 3G/4G femtocells demonstrated at Mobile World Congress next February.
Will the promised nirvana of SON (Self Optimising Networks), where network planners become almost redundant, be achieved?
There is a lot of ongoing development across the industry on SON, much of it benefiting from advances made by the femtocell industry. Doug recently gave an extended (5 hour) presentation on the topic at a recent conference.
Although the industry is generally very conservative, the more technically advanced operators are encouraging its development – they see it as a means to reduce today’s high ongoing costs of radio optimisation while simultaneously increasing network performance and efficiency. There is a growing market need for this feature, which will be essential as the number of small cells increases.
SON won’t completely absolve the need for some manual radio planning activities – for example, some human oversight will be needed to prioritise where and when to install additional basestations. But we have seen a different mindset in operators who already have experience of SON by testing and rolling out femtocells. Other operators disagree, and we will see different approaches taken as the technology becomes more widely available.