Are we clueless on wireless capacity?

At the Institute of Electrical and Electronics Engineers' International Symposium on Personal, Indoor and Mobile Radio Communications in Sydney I organised a panel – ‘Innovative ways of using wireless broadband'.

The conference brought together over 600 leading researchers on wireless technology from around the world. For first time in its 23 year history the event was held in the Southern Hemisphere.

The topic was discussed by the panel members, which included: Professor Luis M Correia from the Technical University of Lisbon, Dr Holger Claussen, head of the Autonomous Networks and Systems Research Department at Bell Labs in Ireland, Professor Jay Guo from CSIRO, and Gary McLaren from NBN Co.

A key theme was that the enormous growth in mobile usage, doubling each year, is set to continue for several years in a row. According to Ericsson, by 2020 mobile operators will need to provide one thousand times the capacity that was required in 2010. The assessment was that the mobile industry has already fallen behind in delivering the capacity needed today, let alone coping with the enormous growth ahead; and that this situation will deteriorate before it improves. A major concern however, is that there is no clear industry road map for the future. There is slightly more clarity in relation to the backhaul technologies, but the end-user access problem is still far from reaching a resolution.

The capacity required for the burgeoning wireless broadband market will need to be addressed in completely different ways. Whatever happens with spectrum allocations, now or in the future, there simply will not be enough of it. Different and innovative solutions will need to be applied – probably involving a combination of solutions and technologies, plus a much higher level of collaboration within the broader industry.

The 30 year old MIMO technology is also being revisited. This technology basically adds a range of antennas to both the transmitter side and the receiver side of the network. It is applied in many industry standards but further developments could assist in increasing the number of antennas, thus increasing the data throughout within the available spectrum. However this theoretical physical solution still needs to be developed into a technological reality.

Fixed mobile conversion

Fixed mobile conversion is another important development. Eventually the backbone network needs to be fibre-based – perhaps as little as one-third of the network has been upgraded to fibre so far. While microwave technology has proved to be extremely resilient with more and more cells now added to the network, the direct line of sight for smaller cells becomes an issue, and this means that fibre is needed very deep into the cities. Some companies, such as Vodafone, are already actively pursuing options to become involved in fibre optic networks (national broadband networks).

The difference between the fixed and the mobile network will become increasingly blurred, with mobile handsets simply being devices wirelessly connected to the fixed network.

All of this will stimulate FMC. Many incumbents already have the option to provide more integrated services but they deploy this in a limited way so as to avoid cannibalisation and maximise their profits.

As BuddeComm has predicted previously, all these developments will eventually lead to a structural separation between the infrastructure and the services. Once operators start to think along these structurally separated business models, solutions to the mobile infrastructure crisis will become much easier to realise.

Small cell networks

This was a recurring topic, not just during the panel discussion, but throughout the conference.

The mobile cells have to be dramatically reduced in size – as small as to cover single rooms, buildings, bus stops, high traffic junctions and so on. This will be a challenge, especially in relation to interconnections (line of sight) interference. An architecture based on smaller cells will also require a high level of intelligence within these cells, as some people will be rapidly changing from one small cell to another. There have been several false starts and so far the results of small cell networks in the USA have not been encouraging. Costs are still too high and there are technical problems still to be overcome – for example, getting bandwidth capacity to these cells in a cost-effective way is already a major issue, and in a booming mobile broadband market the need for more capacity is only going to increase.

Because of its complexity, small cells on their own are not going to solve the capacity problems the industry is facing.

Heterogeneous networks

Another unavoidable development is that a range of technologies, frequencies and protocols will be needed to overcome the problems. CSIRO's Ngara network, initially developed as a mobile access technology using TV spectrum (white spaces), has since moved into the backhaul network and there was no doubt among the panel members that whatever backhaul technology is available and makes sense will need to become a part of such a heterogeneous structure. White space technology does cause problems at the mobile phone level and is therefore not well-suited as a user access technology, but it can be used for backhaul operations.

We already mentioned fibre but WiFi and WiMAX are also already deployed. One solution that operators are using is that mobile broadband traffic is offloaded from the mobile network ASAP. It is estimated that 80 per cent of mobile broadband usage takes place via WiFi connections back into the fixed networks. Over 60 per cent of tablets don't even have a mobile connection. They rely solely on WiFi connections. As we have mentioned before, the enormous growth in demand for mobile and wireless broadband will also lead to a further explosion in WiFi. A more seamless integration of WiFi and cellular technologies was another issue that was discussed at the conference.

Muddling along

With no clear future direction the existing technologies and tools will have to suffice in the meantime, which means a mobile broadband infrastructure shortage for at least the next five years. A side effect of all of this is that all mobile networks have to work with data caps in order to manage the capacity shortage and this makes mobile broadband very expensive compared to fixed broadband. On average, capacity on a mobile network is 5 to 10 times more expensive than capacity on a fixed network.

Among the engineers at the conference there certainly was no doom and gloom; on the contrary, there was excitement and optimism. Also, the majority of these engineers were young people, most of them coming from China, India and other places in Asia, so there is certainly no lack of human resources to work on the problems at hand.

However, at the same time the conference was unable to come up with a clear road map or any clear-cut solutions. Even under the most optimistic scenarios, with more spectrum, more efficient use of the spectrum and new technologies, such as small cells, this will not be sufficient to get ahead of the issue any time soon. What the most likely solution will be also remained unclear and this makes it all the more difficult for network operators to make appropriate investment decisions.

Paul Budde is the managing director of BuddeComm, an independent telecommunications research and consultancy company, which includes 45 national and international researchers in 15 countries.