What type of cellsites are being targeted here?
First lets outline the four main classes of cellsites:
Picocell or Enterprise Femtocell
Typical max active users
RF Power dBm
Today’s femtocell industry has been heavily promoting residential and enterprise use, although some examples of Public Access use have been quoted such as SoftBank Japan (using Ubiquisys) and Vodafone Qatar (using Alcatel-Lucent).
This announcement from picoChip launches a new chipset variant, the PC333, specifically designed for these larger Public Access cellsites provided by operators with small to medium range.
Although we are all familiar with the highly visible celltowers dotted around our landscape, these are not the type being installed most often today to improve coverage and capacity. Mobile phone technology addresses both of these issues by installing many more smaller, unobtrusive basestations in areas of peak usage. These can cover street canyons, transport hubs, campuses and similar hotspots.
The new chipset specification
Today’s announcement by picoChip relates to their PC333 chip, which will be available for sampling during Q4 2010. Products based on this chip are expected to be in customer trials by April/May 2011.
The PC333 is a further development of picoChips recent PC3xx series incorporating features required of this more complex class of basestation. Receive diversity (to allow you to choose two received signals and chose the best, or mix them to get a better answer) is needed for more than a few users.Where residential femtocells don’t need to cope with users driving past in their cars, those covering outdoor streets clearly do – so vehicular mobility at speeds up to 120km/h is included. Soft handover, where single devices can concurrently talk/listen to more than one basestation and select the best signal on a packet by packet basis, is another example. Range exceeds 2km and the design uses receive diversity to handle the difficult and harsh RF environments common in urban areas.
Specifically, the PC333 meets the 3GPP Local Area Basestation (LABS) level of conformance, which has more stringent test cases than the “Home Basestation” class used for most femtocells.
Finally, the PC333 supports picoChip’s smartSignalling feature, which allows them to cope with up to 400 smartphones on one cell, all in “background” mode for push email or apps. This is even more important with larger cellsites, where the signalling load can otherwise clog up capacity.
So is it called a femtocell or picocell?
Some would say that attractive as this approach is, it still picocell or microcell; made cheaper from the use of SoC (system on a chip) technology, but not really changing the architecture. In this case, designs will not use the Iu-h femtocell standard, but instead the existing Iu-b interface common to all 3G cellsites today. Both architectures are suported.
For Iu-b based design, these public access basestations would be installed and managed using the same systems in place today for other outdoor cellsites – they are not dependent on residential femtocells or femtocell gateways being present.
In contrast, others (e.g. Vodafone’s Kenny Graham) very much view this category as part of the femtocell landscape, and build on the capabilities of Iu-h and TR.096. They could be managed using the same systems already in place for their residential femtocells, arguably reducing OpEx as well as CapEx.
picoChip explain why they are taking this route
Rupert Baines, Marketing VP at picoChip, explained this new opportunity to me. “There is a growing consensus that residential femtocell technology works. Traditional basestation vendors (regardless of their views on residential femtocells) are now incorporating this same technology into the designs for their own local area basestations. These benefit from the cost advantages embedded in femtocell designs, but are adapted to meet the higher power and other requirements for public access.
"But to be clear, we will support both Iub picocell and Iuh femtocell: whatever your view of the network architecture, the same PC333 can be used to implement the PHY, and meet the tougher LABS performance level”.
What are the dramatic price implications?
picoChip gave an example of the factory price of a local area microcell basestation at around $20,000 today. Rupert suggests that a similar specification product could be produced using their new design for $2,000 – a reduction of up to 90%. This is achieved through the high integration of features onto a single chipset.
These products will always be more highly priced than residential femtocells. A higher transmit power requires a high specification RF amplifier which might cost $200 or more (for a 24dBm power level) or much more for 30dBm - these are a significant cost component of larger cellsites. Additional protection from the environment such as tolerance of a wider temperate range and robust casing also add to the cost.
Who might lose out if the picoChip design is adopted?
Most basestation designs today use highly configurable but more expensive chips, such as FPGAs. Some allow new versions of 3G to be software downloaded, enabling faster speeds and other features. The consequence of such highly adaptable hardware is higher cost and power consumption, with a very much higher pricepoint.
Not as large a market as residential femtocells?
The forecasts for takeup of residential femtocells is much larger than for these local area basestations, so why is picoChip also encompassing this market? I suspect that although the number of chipsets would be lower, the additional features/capacity and attractiveness to current designs will command a premium price per unit.
Current Analysis Peter Jarich says "... Until now, an HSPA-capable basestation with 32 or 64 channel Release 8 LABS specification would have been called a ‘microcell’ – a market segment that’s largely been ignored by most major 3G basestation vendors. Supporting this type of infrastructure development, picoChip’s PC333 could shake up the 3GPP RAN space – helping to extend the life of 3G as LTE matures". I'm not sure that I agree that the major 3G basestation vendors have ignored the market segment - after all, the industry doesn't buy it's microcells from anyone else these days - but if this announcement does introduce radically reduced pricepoints, then either the major vendors will need to adopt it or risk facing some very price competitive products from new vendors.
Overall, this announcement and likely market takeup by mainstream basestation manufacturers endorses the low cost approach pioneered by the femtocell industry. It reinforces that the “basestation on a chip” approach not only works satisfactorily, but is becoming accepted by the major players.
This is good news for all players in the femtocell industry itself, whether residential, enterprise or public access.