In virtually every episode of Star Trek, irascible engineer Scottie would say, “Ye cannae change the laws of physics.” And those same laws dictate that the electromagnetic spectrum is finite. Whilst we try to push against the laws by using higher and higher frequencies for transmission, at greater cost and greater complexity, the reaction of governments, both nationally and internationally, has generally been to license specific frequencies to specific users. The most extreme examples of this were the “3G” auctions around the world, where mobile phone operators initially bid seriously silly money for the rights to use certain frequencies to operate the then undefined 3G services. In the UK alone, $34 billion was paid by networks to buy access to only 30 MHz of spectrum.
So when a swathe of spectrum becomes available, you would expect that government bean counters would be licking their lips at the thought of more license money, particularly in these straitened times. But although 150 MHz of spectrum has become available, governments, including those of the US and UK, are leaving it free for unlicensed use.
The new “white space” has been released as governments have forced broadcasters to close down their bandwidth-hungry analog television services and use digital transmission instead. A digital network needs only a single narrow band to broadcast simultaneously from multiple transmitters. For analogue, multiple transmitters needed multiple frequencies. Without these the TV receiver could receive multiple copies of the same signal, separated by a time difference depending on how far away the transmitters were. In the 1960s, I lived near a building site with a huge tower crane. At night the crane was left free to turn in the wind, almost like a giant wind vane. Depending on the wind direction, we would get ghost images as a relatively strong second signal, reflected from the crane, arrived a fraction of a second after the first. When set manufacturers started using digital processing within the box, this alleviated the problem, but the spectrum allocation remained.
This white space is clearly valuable, in theory. In practice it is not as simple as it at first appears: the free spectrum varies from area to area – there is not a clear set of frequencies across a wide area. A number of organisations have used this variation to successfully lobby governments and regulatory authorities to make it unlicensed spectrum, like that used for Wifi at 2.4 GHz and 5.2/5.3 GHz. This lobbying has been successful, and the last few weeks have seen several announcements about attempts to develop products and services.
The biggest splash was Microsoft’s launch in Cambridge, England, of a series of field trials of connectivity in the TV white space. Microsoft has been active in lobbying for the license-free use of this space, and, for the trials, it has gathered together a consortium of some very big beasts. They include the BBC, the programme maker and broadcaster; BSkyB, the satellite TV company (Rupert Murdoch’s bid to move from 39% ownership to 100% is running into serious obstacles, and BSkyB, like its rival Virgin Media, is moving into broadband and home phones and has just bought The Cloud, a company providing public Wifi hotspots); BT, the UK’s equivalent to ATT, a previous fixed telephone line monopoly fighting to get a new role in a broadband, wireless and competitive world; Cambridge Consultants, a leader in wireless consultancy and design; Nokia, the mobile phone and mobile phone network technology company; Samsung, the multi-faceted electronics company; TTP, a technology development company based near Cambridge, and Neul, a start-up creating products for white space. We will be coming to Neul in a minute.
The reason for Cambridge is that, apart from Microsoft, Cambridge Consultants, TTP, Neul, and BT having labs there or nearby, it is an interesting area. The city centre is compact, with buildings that are many hundreds of years old and built with walls of thick stone. The dense population is a heavy user of technology, both wired and wireless, placing heavy loads on existing service providers. Yet within a few miles are villages with poor 3G coverage and limited broadband access.
The services that the trials will examine will include using the TV spectrum’s range and ability to penetrate stone walls to extend broadband into the villages, offload data from the 3G network and so improve mobile phone voice coverage, and to provide streaming video and audio from the BBC and BSkyB to hot spots around the city. The trial will also look at the “Internet of things,” the new name for machine to machine communications (M2M) for things like electronic signage, car to car communication, and the linking of other kinds of intelligent devices.
The trials are not going to involve the public at this stage, although there will be demonstrator systems, and people in Cambridge can expect to see all sorts of visitors as the consortium shows off its toys to other companies and regulators from around the world.
It is the Internet of Things that particularly interests Neul. The company was founded in 2010 by a team including founders of CSR (previously Cambridge Software Radio – in itself a spin-out from Cambridge Consultants). CSR is a fabless-semiconductor company that has a dominant position in Bluetooth and other wireless technologies.
The team sees the white space as being the fabric for what it is calling “a truly ubiquitous wide area network,” and it has developed a networking technology specifically for this area, NeulNET. The company is also about to launch a new, open, communication standard for M2M communication, which will be called “Weightless”.
The one-year-old company is already shipping products, and it makes some very interesting claims for its technology. While regulators are freeing up the spectrum, they are putting in place some stringent standards. In the US, the FCC has framed Adjacent Channel Power (ACP) standards, which are designed to stop signals interfering with digital TV signals. Neul claims that NeulNET products meet these specifications. Even with these standards, TV signals may spill over into adjoining frequencies, creating grey space: Neul’s systems are designed to cope with this interference. The free frequencies vary between areas, and NeulNET will have knowledge of what frequencies are available. In both these cases, frequency hopping will minimize interference.
Base stations, which will use conventional Internet connectivity for backhaul, will have “a similar power output to a mobile phone,” yet will have a 10km (6.2 mile) range at 16 Mbps. Tweaking the modulation will allow balancing range and building penetration. A single base station supports up to 1 million M2M connections simultaneously, and a network can support up to 50bn connections. Terminal chipsets will be around $1, as the design is of an equivalent complexity to Wifi or Bluetooth chips. This same design will permit up to 15 years’ battery life for low bandwidth applications such as smart metering. There is also network management software, including knowledge of the spectrum availability for a specific area downloaded to the base stations over the internet connection.
An example application for NeulNET is the shelf-edge pricing labels on supermarket shelves. Even when supermarkets spend a lot of staff time updating the labels, it is still difficult both to react quickly to competitor’s promotions and to make sure that shelf and till prices are synchronized. There are already r.f. and infra-red systems, but the environment of the store makes getting signals to the labels complex: NeulNET should provide a much better solution, with a single base station able to cope with all the labels of even a very large hypermarket.
Two characteristics of an M2M system are that there is often very little upload (if any at all, in the case of the shelf edge labels) and that downloads are not frequent and are usually very short. Modelling a national network, on which service providers would sell bandwidth like today, Neul calculates that the whole of the UK could be covered with 6,000 base stations (currently there are 54,403 mobile phone base stations, with greater London alone having nearly 7,000). Each base station would have a 4km service radius (about 2.5 miles) and, if every man, woman, and child had ten terminals, each base station would support around 100,000 terminals, a tenth of its theoretical capacity, and could send every terminal a 50-byte message every minute. Smart metering is coming in Britain, and less than 1% of the capacity of the network could serve all the British smart metering requirements. (I love it when a company gives lots of numbers like these to play with.)
Neul is shipping base stations and demonstrator terminals today and is in discussions with future terminal developers, who will buy chips or license technology from Neul.
These are wonderful claims, and we have heard similar stories of how other new technologies could transform the world. Zigbee was going to be everywhere, but it is still struggling. However Neul is in with the big boys as a part of the Cambridge trials led by Microsoft, and it is in other trials around the world, which they can’t talk about. The team has obviously convinced the investment community, as they have just raised $12.8 million, in part because of the founders’ record with CSR.
People are queuing up to play in the White Space, and you will see a slew of new applications to exploit the spectrum. The next couple of years could be fun.