At a recent event to promote the UK BSG’s ‘Pipe Dreams’ project, the first question put to the government minister for Competition, Stephen Timms MP, came from a representative of a mobile operator. He questioned the basis of the project’s fixation on fibre. As a self-confessed technology enthusiast, ‘I speak as a geek’, and as one familiar with HSDPA, he could not see evidence of an increased need for even greater speed.
On the same day at an Ofcom/CMA seminar on BT’s 21CN, there was criticism of BT’s websites. Voicing a plea for ‘lite’ versions a delegate pointed out that many of us were increasingly inclined to use mobile devices for Internet access and these websites were, apparently, far too content-rich to be easily accessible via PDA’s.
Between these two contrasting observations lies a tension that demonstrates that with increasingly diverse markets, where customer requirements range from anorexic to obese, one size does not fit all.
Once you strip away the underlying motivations, the heavy burdens of sinking Capex and the race to reduce Opex, the arguments amongst the tribes of telecoms about the relative superiority of their chosen technologies (and their supposed need for regulatory support) become largely irrelevant. What ultimately matters is that the services they provide should (a) deliver something that customers really want and (b) make a profit to repay the investment.
Over the past decade the challenge for investors in advanced mobile technologies has been to identify markets large enough to justify the infrastructure outlay. This challenge, however, is about to ease as investors wake up to proven mobile technologies that can deliver profitable services to what were previously regarded as niche markets. They will do this by making a far better job of meeting specific needs. There is, in effect, a third generation but it is not the 3rd Generation as we currently know it.
The UK’s fixed-location DSL services are limited by line-length but offer around 80% geographic coverage and contended download data rates of between 2 – 8Mb/s. This reality is significantly less than the claimed headline speeds but expectations are still fuelled by the promotion of HD-ready TV kit and media players. The greatest generic broadband shortfall is in rural areas where competitive choice is limited.
The DSL-under-served rural market has long been seen as having potential for wireless broadband (fixed or mobile) but few investors have yet fathomed how to build viable business models that meet customer expectations of affordability and performance set by their suburban relatives. At the same time there has been a proliferation of WiFi schemes, both urban and rural, that attempt to meet the additional needs for occasional Internet or access. In other countries, notably Japan, mobile data access has always played a larger role.
The Japanese market has long been accustomed to an end-user 384kb/s service on the move. This induced a rapid proliferation of mobile usage and, as a result, Japanese technology experience in the use of Adaptive Antennas to make mobile broadband economically viable is unrivalled.
Multiple Antenna Systems (MAS) have now appeared on the European and North American media radar. It has emerged that MAS will be needed for the push to try and make the 2010 mobile versions of WiMAX economically viable – or at least perhaps to begin to catch up on the more-advanced mobile services on view in Japan and in commercial operations in Australia, South Africa, Malaysia, Norway, Canada, USA and several other countries.
So how do MAS work and why do they make a difference?
Figure 1 shows a roof-top in Yokohama with three generations of adaptive antennas.
The oldest, in the middle, uses just two antennas. Introduced in 2000, this design boosted the original Personal Handyphone Service (PHS) to 384kb/s – roughly equivalent to recent European 3G services. It was based on ArrayComm’s original Intellicell® design.
Each of those two antennas is a vertical stack of small dipoles encased in a thin tubular cover – effectively creating a collinear array. PHS may now seem outdated but with handset sales in excess of 54 million it has had a fair run.
On the right is a 2nd generation 8-antenna Super-WLL system - providing 155 concurrent voice channels plus a data capability equal to enhanced 3G services. This Wireless Local Loop design was intended mainly for deployment in areas where wired distribution networks were under-developed.
The 12-antenna iBurst system on the left (introduced in 2003) delivered a fully mobile 1.2Mb/s standard IP data service – and it is this system that has now been refined and commercially deployed in thirteen countries. This design conforms to the HC-SDMA ATIS/ANSI Standard. It is included in ITU-R Recommendation M.1801 for BWA systems and is one of the candidate TDD technologies for the long-delayed IEEE802.20.
At first glance the technology is all about directionality. Take the earliest 2-pronged system and liken it to your own hearing: your ability to sense the direction of a sound and focus on it to the partial exclusion of other noises is not far removed from the processes applied to signals from a user terminal arriving at each dipole at fractionally different times. With this 1st generation MAS there would typically be an array of 2 X 8 dipoles. Signals from a user terminal arriving across this array have a specific phasing and amplitude – fractional timing differences and signal strengths.
The inverse of the pattern arriving from the user terminal is used to direct the outgoing signal energy towards the user and to the exclusion of other areas in the coverage footprint. This is called Spatial Division – effectively creating ‘personal cells’ that adapt to the user’s movement across the coverage area. This means that the same frequency channels can be re-used at the same time in other directions. The result is a massive increase in spectral efficiency, more users and traffic capacity, and far less interference. Hence the codification of this design as High Capacity-Spatial Division Multiple Access, HC-SDMA.
About 5 years ago the 2nd generation MAS took this a step further by increasing the number of antennas and dipoles to improve the degree of accuracy of these ‘personal cells’. This in turn meant that user velocity was also improved –from nomadic pedestrian pace to 30mph. Handset design also kept pace with this improvement – fuelling ever higher user expectations of Voice with added data. The gradual improvement in these technologies owes much to faster processors, better predictive algorithms, and slicker combinations of Spatial and Time Division (SD+TDD) techniques.
The current 3rd generation of this technology (iBurst®) was introduced in 2003. It samples each user’s signal phasing and amplitude 200 times a second – so a user’s velocity can now reach 70-80mph before beating the predicted focus for the returning signal. In practice, as travel velocity increased above 80mph, the downlink performance rate degraded slightly more than the uplink speed.
Spatial Division (SD) techniques combined with TDD management of time slots and sectorisation produces higher levels of spectral efficiency. The WiMAX Forum’s 2006 projections for 2010 performance (Fig. 2) showed the potential impacts of adopting Multiple Antenna Systems to improve mobile performance.
Fig. 2: Source and © : WiMAX Forum: MobileBroadband_personal broadband – 17-08-2006
All manufacturers now expect to improve on these predictions and some claim to have doubled their spectral efficiency. However, the comparable download value for the 2006 production version of HC-SDMA/iBurst was already standing at just over 6bps/Hz, illustrating the advantages of a design that fully exploited these concepts from inception.
Conventional wireless systems design places a large emphasis on elimination of interference. Typically all manner of reflected signals, clutter, get in the way of identifying a clean connection. These MAS, however, thrive on multi-path signals – the more complex the pattern of reflections the more accurately each user terminal is uniquely identified. This becomes particularly important when seeking indoor penetration for mobile terminals that may be in close proximity. To achieve this level of performance the frequencies used must preferably be below 2 GHz – well below those envisaged for WiMAX. The HC-SDMA design also provides significant advantages in ‘interference mitigation’ – avoiding conflicts with close spectral neighbours such as GSM and enabling otherwise unusable spectrum to be brought into commercial use.
Another huge benefit is that systems using MAS can delivery high performance with far less spectrum. The deployment economics of HC-SDMA benefit from needing only a single band of 5MHz compared to the larger paired frequency bands used by older mobile designs and the 21MHz band (plus guard band) needed for WiMAX. It’s been estimated, for example, that with HC-SDMA, mobile broadband data services for the whole of Europe (even assuming 3 competing operators in each country), could manage quite well with less than 5% of all the spectrum currently (and expensively) licensed, allocated or reserved for 3G phone services. The dramatic impact of lower spectrum costs on the investment case was recently illustrated when licences for HC-SDMA deployment were acquired by Personal Broadband for the whole of Ireland (including Northern Ireland) for less than $1m.
Another consequence is that fewer base stations are needed. Like-for-like coverage of, say GSM, with the new technology needs perhaps only 50-60% of the cell sites but can deliver a high-speed IP data end-user service with lower RF radiated power levels. Even more usefully, the system needs none of the conventionally complex and expensive RF planning effort because the same frequencies are used by all base stations. This makes it very easy and far less expensive to add system capacity. In tune with most Telco strategies for migration to all-IP networks (data networks with a voice capability), HC-SDMA is a 21st century mobile IP data system that is a million miles away from legacy circuit-switched voice telephony designs. This makes it possible to deliver a wholesale service that can easily be plugged into ISP and other Communications Provider networks, and that in turns creates new opportunities for retail competition.
Inevitably there are downsides. Nobody has yet packed the electronics for the 3rd generation MAS-based services into a small handset – so for mobile users it’s currently available as a PCMCIA card or USB plug-in for lap-tops and notebook PC’s.
Dual-mode battery-operated WiFi/HC-SDMA modems are beginning to appear and these have been deployed in conjunction with mobile 2G/3Ghandsets also equipped with WiFi – opening up the possibility of using HC-SDMA as mobile backhaul where the user might otherwise have depended on WiFi hotspots.
Some of the earliest users in Australia soon cottoned on to the idea of having their own mobile WiFi hotspots within range, say, of their parked cars. The Sydney Metro trains use the technology for passenger information and advertising updates and there’s even a radio station, Bondi FM, that operates from the beach using iBurst for its studio link.
The good news for the majority of mobile business people is that HC-SDMA cards use less than half the battery power of emergent WiMAX designs - so laptops or notebooks will hold up well in places without mains power. The technology roadmap for higher speeds - both data throughput (8Mb/s) and travel velocity (100mph) - is rooted in better maths and faster processing. Kyocera has this year introduced a new chipset for user terminals that will allow smaller units consuming even less power to come into large-scale production.
To convince a sceptical bunch of journalists who have been long become hardened to industry hype, Kyocera, mounted an impressive demonstration in Yokohama during the 2005 Broadband World Forum. Engineers always shudder when the boss insists on a live demonstration – and this particular high-wire act for technology experts had no safety net.
The hotel chosen for this event was in one of four base-station areas forming a chain across a swathe of Yokohama that included an urban motorway – convenient for testing cell hand-off at high speed. Arranged around the room were 21 lap-tops connected wirelessly only by iBurst. These were clustered in three banks of seven to demonstrate proximity working.
They were taken through their paces simultaneously with tasks of downloading videostreams and web-browsing at 1.2Mb/s. The audience was treated at the same time to a large screen display of the system throughput monitors with sufficient detail to show the loss of one signal when one of the lap-tops was deliberately shielded by covering its antenna with a steel tube. Being modest, none of the presenters pointed out that an aggregate throughput of 24Mb/s contrasted favourably against an original 3G base station capacity of 384kb/s.
Having showed off its conventional IP data access capabilities, attention then turned to VoIP. Using a small IP PBX the demonstration team showed how a single iBurst link could provide access for four concurrent outgoing or incoming VoIP calls. The demo wrapped up with video replays of engineers driving along the urban motorway to observe call handover as they passed from one base station area to another – complete with an on-screen countdown to help the audience catch that moment when we were supposed to be able to detect a faint click at the hand-off point.
Setting aside the theatricality, it was a very convincing demonstration that went way beyond most media expectations. The Kyocera engineering team confirmed their commitment to add carrier-grade QoS for VoIP traffic across the iBurst system and the results of this work is now coming through to commercial deployment.
In truth (even in 2005) nobody was really surprised. The whole affair was a reinforcement of the commercial success already experienced in Australia and South Africa – but Yokohama is the home of Kyocera’s research labs and the engineering team needed to demonstrate the impact of their company’s investment.
Meanwhile in Europe the early demonstration site for HC-SDMA/iBurst in Belfast has now been overtaken by a commercial deployment in Norway and the first contracts for the Northern Ireland roll-out. Development work continues at the Northern Ireland Science Park where the focus is now on bringing together a wide range of industry partners and potential resellers to develop enterprise mobility applications - embedding the service into a myriad of transport, logistics and monitoring systems. With assistance from many local organisations iBurst has been applied on the move to user needs as diverse as a film crew working on an oil rig, pilot boats managing the arrival of large vessels and mobile video surveillance for public safety and security organisations.
Jim Cooney (founder of Personal Broadband and successful bidder for the all-Ireland spectrum, 1785-1805MHz) is sure that the innovation ushered in by iBurst technology will change the way we all think about mobile services. “This is”, he says, “light years on from the business models of established mobile operators. We do not have to bend over backwards to accommodate IP data or replace legacy kit to stay in the game.
Our challenge, along with our Channel Partners, is to satisfy the market for robust and secure mobile IP data access. We have the huge advantage that our IP-based Spatial Division approach was always a ‘next generation’ system and, from an investor viewpoint, it’s one that has already proven to be commercially viable.”
The scope for mobilising ‘information intensive individuals’ – e.g. health care visitors, architects, construction managers, local government planning inspectors and emergency services personal – is obviously vast – but the needs of these professional users are very different to the massed ranks of mobile consumers. It’s also becoming obvious that the security issues associated with WiFi hot-spots and ‘guest access’ often conflict with the need for convenient mobile access to secure VPN’s – a fairly vital part of any corporate requirement for greater ‘flexible working’.
This suggests a wide gap between different market perceptions. The apparent Telco interest in WiMAX, for example, is still seen as an opportunity to plug gaps, the ‘not-spots’, in their fixed broadband coverage for consumers. The mass market dreams of delivering customised HD IPTV streams over vast distances within the limitations of available spectrum still baffles investors and engineers.
After dotcom boom, bust and excessive wireless industry hype, it has to be good news that investors are still willing to back innovative services for selected markets. The clarification of a real market for robust enterprise mobility is just part of the story. The other part is the existence of a commercially proven mobile technology that actually delivers what it promises – the next generation of secure personal broadband.
__________________The author, David Brunnen, is MD of ABFL Groupe Intellex, a consultancy practice specialising in the commercial development of ‘disruptive technologies’. This article is a revised and updated version for Land Mobile magazine, October 2007. An earlier version of this article first appeared in 2005.