A Way Forward for Broadband Wireless Access An Investigation for The Radiocommunications Agency (March 1999)

This study has clearly shown that both potential operators and equipment manufacturers believe there is a real market demand for broadband services and that Broadband Wireless Access is now a sufficiently mature technology to satisfy this demand. The impetus comes initially from the need of small and medium sized businesses for higher speed data connections but this will be followed quickly by new applications such as collaborative working and peer to peer video telephony. Within a few years, Broadband Wireless Access will also bring broadband services to the residential sector.

The growing convergence of the technologies of voice, data and video entertainment is rapidly leading to the need for flexible broadband communications direct to the small business and to the home. Upgrading the copper based local loop to provide this capability will come but will be both slow and expensive. In these circumstances, wireless networks have a number of key advantages. In particular, they are less expensive and can be installed within substantially shorter time scales than their wired counterparts.

These proven radio benefits can now be applied to the delivery of broadband services on a local, regional or national basis. Broadband Wireless Access is therefore an ideal solution, and will allow new and innovative operators to provide a competitive service from day one.

Initially, two types of operator are likely to emerge, those wishing to offer broadband telecommunication services primarily to businesses, and Cable TV companies wishing to offer telecommunication services in addition to TV in areas where it would be uneconomic to install a fibre network. Thereafter, both types of operator will expand their range of services until they become full service access network operators providing a complete range of services from telephony through interactive multimedia to high definition TV. In parallel, other operators may emerge to build new wireless based full service access networks.

The European community clearly foresee this to be an important market with ETSI establishing the Broadband Radio Access Network (BRAN) Project Team in 1997. The general view of respondents was that licensing should not be delayed unduly by the standardisation process while adequate ETSI standards already exist or are being developed which ensure that the UK will not differ in its approach from the rest of Europe. Cognizance should also be taken of the advantages that a prompt start would offer UK companies in export potential both for equipment supply and system design.

The study suggests that there would be advantages deriving from offering licences on a regional rather than a national basis and that each region could support two licensed operators and a number of local operators. It further suggests that the spectrum should be made available in tranches with the first wave limited to a single operator per region.

At the present time adequate technologies exist in a proven state to address these service requirements. A significant number of companies now have products available which offer end user connections at the 2 to 8 Mbits per second level and a number of trials are currently under way. The majority of potential operators believe that 2 to 8 Mbits per second will satisfy most small business requirements in the early phases of the market. Within 5 to 7 years, applications for the small business sector will require bandwidths of 8 to 34 Mbits per second, and residential subscribers could be making use of services at 2 to 10 Mbits per second.

The obvious choices are the 28 and 40 GHz bands. At 40 GHz, the 40.5 to 42.5 GHz band has been made available to the ITC for use by LDO licence holders for Multipoint Video Distribution Systems (MVDS) but this spectrum may be available for Broadband Wireless Access where it is not being utilised by LDO licensees. It is also possible that the 27.5 to 29.5 GHz and the 42.5 to 43.5 GHz bands (net of sharing issues) could be made available for Broadband Wireless Access and any supporting point-to-point links. In this context, the World Radio Conference 1997 (WRC ‘97) agreed a Fixed Service allocation alongside Broadcasting in the 40.5 to 42.5 GHz band and also made available spectrum in the 31 GHz band.

The clear view of the majority of respondents to this study is that initial licensing should take place in the 28 GHz band.

Over the past few years, there has been an increasing interest in wireless as a means of providing broadband connections to business and residential subscribers. Licences for Broadband Wireless Access were issued in Canada during 1997 and similar licences were auctioned in the USA in March 1998. Applications for broadband wireless access licences at 26 GHz are currently being considered by the German authorities and licences to provide similar services were auctioned by the Australian government in February 1999. At the same time, the Radiocommunications Agency has received a number of submissions to the effect that the UK market is ready, or close to being ready, for Broadband Wireless Access.

In response, the Agency commissioned Quotient Communications to undertake a brief study into the potential demand for Broadband Wireless Access. The objectives of this work were:

For the purposes of this study, broadband services were defined as those which required a connection to the individual subscriber at 2 Mbits per second or faster.

It was intended that this study would inform the needs, scope and content of a Consultation paper provided the level of interest justified it. The majority view of respondents was that, given the market readiness and the level of technology development, it was now appropriate for the Government to move forward with a view to issuing licences as soon as possible.

Early on in the project, it became clear that there were a wider range of interests with a potential involvement in Broadband Wireless Access than had been expected. We therefore modified our original approach to put more effort into discussions with all sides of industry. Despite this change of tack, the scale of the project precluded a very comprehensive dialogue with industry and we are aware that some interested parties have not had the opportunity to contribute their view-point. We trust that this will be redressed during any public consultation.

The results presented here are based very largely on discussions with interested parties, supplemented by a limited amount of desk research. The study was originally completed in February 1998. A further brief study was undertaken at the beginning of 1999 to identify any significant changes relevant to the UK that have taken place over the past 12 months, and the results of this more recent work have been incorporated into this report.

It is clear from the preceding comments that the contributions of industry to this study were vital, and the Radiocommunications Agency and Quotient Communications Limited wish to thank all those listed below who were generous with their time and knowledge.

The following chapter, Chapter 3, provides a brief introduction to Broadband Wireless Access and how it fits into current developments within the telecommunications arena in general and into the UK situation in particular.

Chapter 4 summarises the findings of this study as a series of topics such as market size. Each topic is summarised and the implications relevant to the regulatory regime and to the Agency’s questions are given.

In Chapter 5 we present an initial set of conclusions which follow, with a degree of logic, from the evidence and implications in Chapter 4. They will require at least some refinement, and much additional detail. They are presented as a starting point for discussion and as a stimulus to debate.

In the traditional telecommunications network, most subscribers are connected to the local telephone exchange via copper wires. Larger businesses, with higher volumes of traffic may be connected via fibre or point-to-point microwave links. Together, this network of connections is referred to as the local access network. Wireless access networks, in essence, are simply an alternative means of making the connection between subscriber premises and the local exchange.

The majority of wireless access networks support traditional telephony and relatively low speed data (for example, up to 64 kbits per second), and are usually referred to as Wireless in the Local Loop. Broadband Wireless Access networks are capable of carrying much higher bit rates (several Mbits per second or more) and, because they are also digital, they have the potential to transport a complete range of electronic traffic including telephony, high speed data, TV and multimedia.

Figure 3.1 In this figure, the local Broadband Wireless Access operator has a radio base station installed on a central building, and business premises in the neighbourhood are equipped with broadband wireless subscriber equipment. In the example shown, one business is using a high bit rate service and has been assigned a wide band channel. Two other businesses are using lower rate services and have been assigned narrower channels. When they finish using the connection, the channel is released and can be used by any other subscriber.

Figure 3.1 illustrates a typical Broadband Wireless Access network. A central radio base station communicates with equipment located in a multiplicity of subscriber premises in the surrounding area. For this reason such systems are often referred to as point-to-multipoint systems. (In contrast, a fixed point-to-point wireless system provides a communication path between just two fixed points.)

When a subscriber makes a call, the wireless control system allocates a radio channel to that subscriber but only for the duration of the call. As soon as the call is completed, the channel is free for use by any other subscriber. Moreover, the system will allocate just enough RF resources to accommodate the call. Thus a narrow band channel would be allocated to a telephone call but a high speed remote LAN connection would be allocated a wide band channel. By sharing the radio resource between subscribers and using only the minimum needed for any particular call, Broadband Wireless Access systems can utilise the spectrum more efficiently than would be the case if each subscriber were provided with a fixed point-to-point wireless connection, and can do so at a lower price.

Modern society is increasingly reliant on the transport and processing of electronic information of all types including telephony, data traffic, and TV and radio. Today, much of this information is treated separately; for example, telephony and TV are usually delivered to homes through independent networks. Increasingly though, all forms of electronic information are being converted into digital format largely removing the distinction between them and making it possible for a single network to handle all types. At the same time, there is a continuing demand to transport more information more quickly. The result is a need for communication links of ever greater bandwidth.

Modern telecommunication networks are already largely digital and capable of transporting high bandwidth traffic. The major bottle neck is the copper based local loop to which most business and residential customers are connected. For the Information Society to develop, the local loop needs to be transformed into the Full Services Access Network capable of supporting a full range of broad and narrow band services from telephony through interactive multimedia to high definition TV.

Upgrading existing copper networks to a broadband capability or replacing it with fibre involves very substantial investments, and takes time. Using wireless as the delivery mechanism offers several advantages. The key ones are:

The UK has led Europe and much of the world along the path of telecommunications liberalisation and now has a very competitive market particularly in national and international services. Because of the high cost of implementing access networks, competition has been slower to develop in the local loop, but even here the UK has managed to engender an increasingly competitive environment. As of the beginning of 1997, approximately 25% of all UK homes had the option to take telephony from their local Cable TV company. This proportion is expected to rise to 75% by 2005. The Cable TV companies also offer telephony to business customers.(Cable TV and Telecoms Handbook 1996)

Wireless is also playing a part. Five Wireless Local Loop licences for the provision of telephony have been issued (to Atlantic Telecom, British Telecommunications, Ionica/ScottishTelecom, Liberty and Radiotel). Two of these were operational in 1997 and a third commenced operation during 1998. However, one, Ionica, went into receivership in early 1999.

In 1996, the Government announced the award of three additional wireless local loop licences at 10 GHz to Mercury Telecommunications (now Cable and Wireless Communications), Ionica/ScottishTelecom and NTL, although the formal awarding of licences was delayed pending the resolution of interference problems. Although 10 GHz equipment can support broadband services, the spectrum allocated to these operators limits the extent to which they will be able to do so.

The next logical step would be the development of competitive broadband access networks. If the demand from the market is there and operators are ready to invest, as this study suggests, then now could be an opportune time for the Government to take a lead, to the potential advantage of the UK economy and industry.

Perhaps the first question to answer is: Is there sufficient interest within the business community to justify a government consultation exercise on the possible licensing of spectrum for Broadband Wireless Access? Our answer is a definite Yes.

It is true that contributors to this study were invited to participate on the basis that they were likely to have an interest in Broadband Wireless Access and, to this extent, the answer is unsurprising. However, the contributors represented a range of interests including telecommunications and entertainment, current and potential network operators, and manufacturers. The great majority are enthusiastic about the potential market. Many are clearly prepared to make substantial investments in this area, and a number have already made initial investments. As a result, we firmly believe that a consultation exercise on this subject will receive a significant and positive response.

Today, telecommunications and broadcast entertainment services are separate industries with their own wired and wireless distribution networks. However, their respective technologies are converging rapidly and it is to be expected that, in the not too distant future, operators will deliver a complete range of telecommunications, information and entertainment services over a single network.

Industry players with an interest in Broadband Wireless Access can be divided, roughly, into three categories: current and potential telecommunications operators initially targeting businesses; Cable TV franchisees with a largely residential market, and a third group who envisage broadband access networks supporting both telecommunication and entertainment services from the start. For convenience, we refer to the third group as Full Services Access operators, although all three groups aspire to become full service operators in the longer term. The aspirations and target markets for each of the three groups are briefly described below.

The most immediate opportunity for Broadband Wireless Access is for the provision of business telecommunication services. Many large organisations already have broadband connection to a telecommunications operator via fibre, cable or fixed point-to-point wireless links. Several contributors to this study, however, are confident that there is a growing market amongst Small and Medium sized Enterprises (SMEs), and that a shared wireless point-to-multipoint system is a cost effective way to satisfy this demand. There is also some interest in offering large companies very high bit rates services (at 34 to 155 Mbits per second) using wireless point-to-multipoint systems. Subject to suitable progress on licensing, commercial deployment could take place in late 1999.

No single application is seen to be the driver in this market. Most contributors believe that it will be necessary to offer a range of services and that it will be important to have the flexibility to adapt quickly to new requirements as the market develops. Some possible applications are listed below:

We note that, although initial applications are likely to generate symmetric two-way traffic, there is a growing consensus that traffic is likely to quickly become significantly asymmetric. For example, in recent trials by Telia, 25 Mbits per second was provided on the down link and 3 Mbits per second on the up link.

Initially, broadband services will be relatively expensive with the result that early adopters will be limited to the more visionary SMEs. As a consequence, all the plans of which we are aware propose to start by providing coverage only in areas where there is a conglomeration of businesses (such as in town centres and business parks) and where fibre or cable networks are limited in their extent. For this reason, most contributors suggest that the regulatory regime should not impose any coverage requirements on licence holders during the early phases of market development.

As this market develops, it is to be expected that the range of services will widen, that prices will fall, and that coverage will expand. In the longer term, many expect that Broadband Wireless Access networks will reach into residential districts and provide a unified telecommunication/entertainment service.

Figure 4.1 illustrates how this scenario might develop in terms of time scales and required bit rates.

Today, Cable TV companies provide broadcast entertainment and telephony services in almost equal proportions although the two are generally treated as two quite separate services. Through digitisation and the introduction of cable modems, however, the cable networks have the potential to support a very wide range of services. These could extend from one way and asymmetric two way entertainment services (such as Video On Demand) through transaction services to two way broadband services (such as video telephony). Cable modems are expected to be deployed during 1999 and the Cable TV companies must, therefore, be considered potentially major players in the any future broadband access market.

Approximately 25% of current cable TV franchises are operated under licences which permit the use of wireless delivery (and the band 40.5 to 42.5 GHz has been allocated for this purpose within the UK). At present, there is no use of wireless delivery although a small number of trials have been undertaken.

Wireless delivery is most likely to be used in smaller towns and villages where the probable number of subscribers cannot justify the cost of a cable network. It is also possible that future cable network implementations might make use of wireless in larger urban and suburban areas. Nevertheless, the vast majority of cable subscribers will be serviced over cable networks and wireless delivery will, therefore, have to mimic the cable network and support the same services.

Clearly, much of this traffic is heavily asymmetric with the larger bandwidths required on the down link direction.

In the longer term, it is expected that new services, such as tele-shopping and video telephony, will require a greater degree of flexibility and full two way capability. Consequently, cable TV networks will evolve to Full Services access networks. The current wireless delivery specifications (MPT 1550 and 1560) do not provide this capability and it has been suggested that the band 42.5 to 43.5 GHz could be used, in conjunction with the 40.5 to 42.5 GHz band, for this purpose.

Figure 4.2 shows how the services provided by the Cable TV companies might evolve with time.

The great majority of contributors to this study agreed that telecommunication and entertainment services will be delivered over common broadband access networks at some future date. Most agreed that the boundary between the two will become increasingly blurred and that new services will be developed which span both classifications just as the Internet is beginning to do today. The small group that we have identified as Full Services Access operators, believe that this convergence will occur within a very few years and lead rapidly to a mass market.

There was a general consensus that low cost subscriber equipment is the key requisite in the mass market for broadband access. For this market to develop, it is believed that subscriber equipment prices of �200 to �400 for the transceiver and any set top box need to be achieved. A number of parties are of the view that with new technology and innovative network design, these prices can be achieved and networks implemented within about 3 years.

We are not privy to much of the detail of such proposals. We have, however, seen enough to convince us that these claims should be treated seriously and taken into consideration in developing an appropriate regulatory regime.

Such networks could be capable of handling highly asymmetric services, such as video on demand, as well symmetric broadband applications such as video telephony and computer to computer communication at burst rates of up to 25 Mbits per second. Again, no one application is expected to be the market driver and it will be necessary to offer a range of services. In addition, it is expected that the availability of broadband access will lead to the development of new services and new types of service and content providers. The network operators will therefore need both the technical capability and the commercial freedom to respond to the market as it develops. The target markets will be both residential subscribers and SMEs.

The immediate driver for Broadband Wireless Access is broadband telecommunication services for SMEs. For the first few years, business and residential services may well be provided by different operators and, initially, service will be limited to a relatively small number of discrete areas. However, there will be an increasing overlap between the two and both are likely to evolve into full service providers.

Although there is much confidence that there is a market for Broadband Wireless Access there is much less certainty as to which applications and what bandwidths will be required, and how quickly the market will evolve. The consensus view is that operators will have to be commercially very flexible and that the technology will also need to be as flexible as possible. This suggests that the regulatory regime should be technology and service neutral and certainly that it should not unnecessarily prescribe the limits of the Broadband Wireless Access market.

The Broadband Wireless Access market is in the very earliest stages of development and market forecasts are therefore particularly difficult to make. This is reflected in the range of estimates for the numbers of subscribers provided by contributors to this study which are summarised in Table 4.1.

	2000	2005	2010
Business subscribers	10K - 200K	200K - 500K	~1M
Residential subscribers	70K - 250K	1M - 5M	2M+

A 1997 report by Ovum Limited (Broadband Wireless: Market Strategies) provides forecasts of the Broadband Wireless market in Europe. Their forecasts, adjusted to the UK, are in line with the lower end of the forecasts in the above table.

The figures given in Table 4.1 exclude Cable TV customers connected via wireless (MVDS). The Ovum forecasts suggest that such customers could number 100,000 by 2000, and 250,000 by 2005.

There is a degree of uncertainty as to size of the market and how quickly it will develop. It would be prudent, in our view, for the Government to consider carefully the way in which the spectrum is made available including the possibility of a staged approach.

Broadband Wireless Access systems are currently being deployed for commercial operation in North America by Teligent, WinStar, NextLink, Formus and others. In addition, a number of trials have taken place in the UK (at 10 GHz), and in Europe (at 26/28 GHz). Table 4.2 lists products from three suppliers and is illustrative of the equipment that is likely to be deployed during 1999.

	Bosch	Ericsson	Netro
Product name	Digital Multipoint System	Airline	Airstar
Multi-access/ duplex mode	FDMA/FDD	FDMA/FDD	TDMA/FDD
Capacity available per subscriber	64 kbits/s to 6 x 2 Mbits/s	64 kbits/s to 2 Mbits/s	64 kbits/s to 8 Mbits/s
Subscriber interfaces	Fractional E1, E1, ISDN BRA & PRA	POTS, Fractional E1, E1, ISDN BRA & PRA, Frame relay	POTS, Fractional E1, E1, ISDN BRA & PRA, Frame relay
Maximum capacity per carrier	67.2 Mbits/s on 2 x 30 MHz	67.2 Mbits/s on 2 x 28 MHz	16 Mbits/s on 2 x 3.5 MHz
Modulation	QPSK, 8-TCM, 16-TCM	QPSK, 8-TCM, 16-TCM	4-QAM, 16-QAM
Frequencies	10 GHz (ETSI), 24.5 - 26.5 GHz (ITU-R), 27.5 - 29.5 GHz (FCC)	10 GHz (UK), 24.5 - 26.5 GHz (ITU-R)	10 GHz (UK), 24 GHz (USA), 26 GHz (ETSI)
Other		Migration path to ATM	Packet switched with full dynamic capability

To date there have been few commercial deployments of analogue MVDS, and none of digital MVDS. There have, however, been a number of trials within Europe. In the UK, the DTI, Eurobell and other partners have trialled two way services over MVDS under the auspices of the Cellular Radio Access for Broadband Services project (CRABS). This project examined how interactive TV, multimedia and telecommunication services could be combined with the basic MVDS system, and how a cellular architecture might be used to provide contiguous coverage. Another ACTS project, Cabsinet is also researching two way operation with MVDS.(Ovum p197)

We would expect that digital MVDS equipment could be deployed commercially within a short period should an operator wish to do so. It is our understanding, however, that further work is required before equipment supporting a return path can be realised commercially.

The products considered so far are all based on the cellular concept in which a base station provides a radio link to multiple premises within its intended coverage area. At the frequencies of interest, line of sight transmission is necessary with the result that it is likely that a substantial number of premises within the intended coverage area will be unable to communicate with the base station. One solution is to overlap the coverage areas of neighbouring cells so that a proportion of the premises are potentially illuminated from two different directions. Figure 4.4 illustrates the principle.

An alternative solution that has been proposed is to use a mesh network architecture. In this architecture every customer’s premises is linked to one or more other premises via a narrow beam point-to-point microwave link. The wireless equipment at each customer’s premises acts both as the transceiver for the customer and as a link in a chain to other premises. The principle is illustrated in figure 4.5. With a reasonable density of such nodes there is a high probability that a line of sight path can be found from an existing node to any new customer.

A number of issues, such as network management, remain to be fully addressed for large scale mesh networks. Nevertheless, we understand that such networks could be deployed within 2 to 3 years.

Broadband wireless access equipment, suited to the delivery of telecommunication services, is available commercially. MVDS systems with a two way capability are being developed and could be available commercially albeit it on a somewhat longer time scale and further development of this technology is to be expected.

In our view there is scope for significant further development of Broadband Wireless Access technology. This suggests that the regulatory regime should not unnecessarily prevent advantage being taken of any such future developments.

The level of coverage and the frequency re-use that can be achieved at the frequencies of interest is currently unclear and the subject of research (for example, within the CRABS project). Further information will become available as trials are completed and networks enter commercial service. Given this and the uncertainty in the services that will be taken up, any estimate of spectrum requirements has to be approximate.

The general view within the industry is that approximately 1 GHz of spectrum will be required to provide broadband services. Our own simple estimates for 28 GHz band indicate that this would be enough to support the high bit rate services and penetration levels expected in a developed market within around 5 years, and would enable an operator to provide contiguous coverage over a wide area. At 40 GHz, twice as much spectrum might be required. It should be noted, however, that these estimates could be wrong by a factor of two or possibly more. The amount of spectrum required in the early phases of the market is likely to be less; a figure of 2 x 112 MHz at 28 GHz has been suggested. Note that these figures exclude any fixed point-to-point links required in the supporting infrastructure.

In Canada, licences for broadband access services were made available in 1996 and provided operators with 500 MHz of spectrum (at 28 GHz) in any one area. Supporting point-to-point links are to be accommodated within these frequencies and additional spectrum is to be licensed at a later date. In the USA, the FCC has auctioned two licences in each of the 493 Basic Trading Areas. One licence provides a total of 1150 MHz at 28 and 31 GHz (split into one 850 MHz and two 150 MHz blocks) and the other provides 150 MHz (as two 75 MHz blocks) at 31 GHz. In Germany, spectrum is currently being licensed at 26 GHz for point-to-multipoint applications in smaller blocks of 2 x 28 MHz.

The above estimates are all based on a cellular architecture. Spectrum estimation for a mesh network (see section 4.4) is more complex and requires network modelling. Such modelling studies have been carried out by a number of parties. We understand that the results suggest a bandwidth requirement at 40 GHz of 1 GHz per operator in a developed market, and about 500 MHz in the initial phase. Note, however, that a mesh network is its own distribution network and the requirement for a supporting network of conventional point-to-point links could be small.

Broadband Wireless Access operators will require significant amounts of spectrum. 250 to 500 MHz per operator is likely to be sufficient initially with around 1 GHz being required in the longer term. Additional spectrum will be needed for any supporting fixed point-to-point links.

Note that mesh networks may be able to provide the same coverage as cellular networks with less spectrum and may have little or no need for a supporting infrastructure of conventional point-to-point links (see the following section).

Within the UK, point-to-point links have been used extensively in the support infrastructure for both mobile cellular and wireless local loop networks. They have proved to be very cost effective and have contributed to fast network roll out. For the same reasons, we would expect point-to-point links to be attractive to Broadband Wireless Access operators.

However, even if point-to-point links are only used for local interconnection of base stations (in a cellular network), the amount of spectrum required is likely to be large. The simplistic examples illustrated in Figure 4.6 (on the following page) suggest that the point-to-point requirement could be between one half and two and a half times the point-to-multipoint requirement.

It can be argued that a mesh network is its own distribution network and consequently that there is little or no requirement for point-to-point links. The degree to which this would be the case depends on details of the mesh network and on the local extent of fibre or cable based points of interconnection. At this stage, we suggest that it would be prudent to assume that a substantial amount of point-to-point spectrum could be required to support any Broadband Wireless Access network.

Clearly, back haul circuits from a concentration point to a switch, whether carried by wireless or fibre, will require very large bandwidths (10’s of Gbits per second).

It should be assumed that operators will wish to utilise point-to-point links and that a substantial amount of spectrum, of the same order as for the broadband access frequencies, could be required for this purpose.

Figure 4.6 Two configurations are shown in which three broadband point-to-multipoint base stations are connected via microwave links to a fourth having a fibre back haul connection. In both cases, each base station is assumed to require a total of 500 MHz to support its traffic. If the point-to-point links use a modulation scheme twice as efficient as the base stations, 250 MHz would be needed to carry the traffic for one base station. Assuming a frequency re-use of 2, the daisy chain configuration would require a total of 1250 MHz (500 + 750 MHz) of point-to-point spectrum. The star configuration could potentially suffice with 250 MHz (but a similar adjacent configuration could well need a different frequency).

As discussed above, Broadband Wireless Access operators are, in the long run, likely to require around 1 GHz of spectrum each. The bands where such bandwidths are available will be limited. The Agency has indicated that suitable bands are most likely to be found at:

In addition, WRC ‘97 made a new allocation of 31.8 to 33.4 GHz to high density fixed services. This allocation is to be reviewed at WRC ‘99 following sharing studies, and will not be available before 2001.

This band is allocated to Fixed Satellite Services (up link) and to Fixed Services on a co-primary basis. At present, the Agency is not assigning fixed point-to-point links in this band but the continuing demand for point-to-point links may result in pressure to do so.

Decisions made at WRC ‘97 make it easier for Low Earth Orbit satellite operators, such as Teledesic, to utilise this band. As Teledesic envisages the use of very large numbers of multimedia terminals whose location would be uncontrolled, sharing with other services could be problematical. Figure 4.7 shows where the Teledesic frequencies would fall.

Figure 4.7 The relationship of the Teledesic up link band is shown relative to the Fixed Services band plan in the 27.5 to 29.5 GHz band.

Should sharing not be possible, the frequencies available would be restricted to a total of 1.4 GHz made up as follows:

2 x 336 MHz;
560 MHz of unpaired spectrum in the lower band;
43.5 plus 16.5 MHz of unpaired spectrum in the upper band;
112 MHz centre gap.

This band is allocated to Broadcast and Fixed Services on a co-primary basis. In the UK and other countries within Europe, the band has been allocated for the microwave distribution of TV (usually referred to as Multipoint Video Distribution Systems, MVDS). UK standards for analogue MVDS (MPT 1550) and digital MVDS (MPT 1560) have been developed, and may be adopted as European standards. The digital standard identifies frequencies for use as a return path but is by no means a complete two way specification.

UK Cable TV companies are licensed under either PDSL or LDO licences. PDSL licensees are constrained to use cable but LDO licensees may also use wireless in the MVDS band (provided they obtain a Wireless Telegraphy licence). Over 80% of homes are within areas already franchised under either a PDSL or an LDO licence. Current LDO licensees account for 20% of franchises awarded and just under 25% of franchised homes. Approximately half of these homes are in franchises in which building of cable based systems has either commenced or been completed. Taken together these figures imply that the vast majority of cable subscribers will be served by cable rather than wireless in the foreseeable future.

During WRC ‘97, this band was allocated to Fixed Services on a co-primary basis with other services (including Fixed Satellite Services and Radio Astronomy). There is no allocation to Broadcast Services.

There are allocations to Fixed Services of between 1 and 3 GHz at frequencies between 50 and 66 GHz. The technology at these frequencies, however, is likely to be significantly more expensive than at 20 to 40 GHz for some time to come. We take the view, therefore, that it is not appropriate to consider these bands for Broadband Wireless Access at the present time.

The amount of spectrum which might be available for Broadband Wireless Access in the short term is limited to 1.4 or 2 GHz at 28 GHz (dependent on the possibility of sharing with Teledesic), and to 1 GHz at 43 GHz. The 2 GHz of MVDS spectrum will not be used by the Cable TV companies over a significant proportion of the UK, and could be considered for use by Broadband Wireless Access. In the longer term, further spectrum at 31 GHz and above 50 GHz may be available.

The limited amount of spectrum identified above, and the bandwidths likely to be needed by Broadband Wireless Access operators, suggest that the regulatory regime should encourage efficient use of the spectrum. In view of this, there appears to be no compelling reason to give the Cable TV companies any priority in access to the 43 GHz band. It further suggests that it would be prudent to reserve spectrum at higher frequencies (such as 60 GHz) for future broadband services.

It has been suggested that coverage obligations would not be consistent with the commercial situation, at least during the initial phase of market development (see section 4.2.). Whilst this may be true, it will be important that the regulatory regime ensures that licensed spectrum is properly utilised.

Broadband Wireless Access will not be the only media providing broadband business or residential services. Indeed, some industry forecasts suggest that wireless may be the minority broadband access technology. (Ovum private communication)The two major competitive media will be Digital Subscriber Line technology and cable networks.

The major competition will come from the use of Digital Subscriber Line technologies (xDSL) which will provide symmetrical and asymmetrical, high bit rate services over much of the existing copper based access network. The incumbent PTOs around the world clearly have a major interest in ensuring that this technology is available and a considerable amount of work is being devoted to this end. For example, the Full Services Access Network group (FSAN, a grouping of largely European PTOs and suppliers) have been developing a common set of access network architectures and interfaces, and a number of manufacturers have already provided equipment for trials. British Telecom, and other incumbent PTOs, are already using High bit rate DSL to provide 2 Mbits per second leased lines. Asymmetric DSL is currently being trialled within the UK and first deployments are expected during 1999.(one of the Local Loop reports)

Very high speed DSL promises to provide subscriber bit rates of 12 Mbits per second downstream and 2 Mbits per second upstream.(FSAN conf, June 1996 and FibreSystems Nov/Dec 97 p27, Communications International Nov 97 p75)

The majority of cable TV networks in the UK are relatively new and could be modified to support cable modems to provide a two way capability of up to 10 Mbits per second. A number of trials have taken place, and others are planned. CWC, NTL and Telewest plan commercial deployment before the end of this year (1999) and an installed base of 200,000 units is forecast by 2002.

Two companies, Easynet and Satellite Media Services, plan to launch satellite based, high speed Internet services in the early part of 1999 at speeds of up to 2 to 3 Mbits per second (on the down link). Later in 1999, British Interactive Broadcasting is expected to offer a mixture of two way data and TV services also via satellite. Terrestrial digital TV broadcasters could also offer similar a service. Whilst these services are likely to be limited in bandwidth by comparison with the access technologies discussed above (particularly on the up link) there will be a degree of overlap in the services offered.

It should also be borne in mind that the 10 GHz licensees could offer services which would overlap with some of the services envisaged for Broadband Wireless Access. In addition, Power Line Technology may give the electricity supply companies a means of entering the broadband access market.

In the slightly longer term, broadband multimedia satellite services have been proposed. These include Teledesic and Skybridge who plan to commence service around 2001/2002. (Local Loop report April 1997, p12)

Clearly, Broadband Wireless Access will be faced with at least one, probably two and possibly more, competing wired broadband access operators in most parts of the UK. From a competition point of view, this suggests that adequate competition might be achieved through licensing no more than two Broadband Wireless Access operators in any one area. (We would add, however, that we have not considered the question of competition in any depth.)

The schematic shown in Figure 4.8, on the following page, gives the four key interface standards of interest to the operator of a wireless based access network (other standards are important but of less relevance to this discussion). Open standards at these four points would give the operator the opportunity to purchase equipment from multiple suppliers. Note that the wireless standards (interface #3) can be separated into two parts; harmonised frequency bands and channel plans, and equipment standards.

There are few existing standards relevant to Broadband Wireless Access although, of course, some existing subscriber interfaces will be supported by broadband networks (interface #1).

On the wireless side, there is a CEPT channel plan for the 27.5 to 29.5 GHz band which would accommodate some of the broadband wireless equipment currently on offer, and ETSI is currently completing point-to-multipoint equipment standards for this band. The channel plan could be disrupted by the Teledesic system (see figure 4.7) and these standards may not be appropriate to some of the mesh networks which have been proposed (see section 4.4).

There are also channel plans for both analogue and digital systems in the 40.5 to 42.5 GHz (MVDS) band. These are the UK MPT 1550 and 1560 specifications.

The European Telecommunications Standards Institute (ETSI) has established the Broadband Radio Access Networks (BRAN) project. It will address wireless access systems with bit rates up to 25 Mbits per second for both business and residential applications. It is expected to develop wireless specifications (interface #3, both channel plans and equipment specifications), and to identify the necessary interworking functions (interface #4). The work is intended to be complete by the year 2000.

The Digital Audio Visual Council (DAVIC) is developing end to end standards for implementing interactive multimedia services, and the FSAN group is promoting a standard User Network Interface unit to interface between wire-based broadband networks and a user’s equipment. This work may provide standards relevant to interfaces #2 and #1. In addition, the Digital Video Broadcast project (DVB) within ETSI has published ETS 300 748, an MVDS downstream digital standard (interfaces #1 and #2) which is compatible with the MPT 1560 specification (interface #3). The LMDS interaction (return) channel standard (interfaces #1 and #2) is currently in draft form within the DVB project.

All contributors were in favour of open standards but the great majority believed that the current lack of Broadband Wireless Access standards should not be allowed to hold back licensing and the development of the market.

The consensus is that open standards are beneficial. However, although there is much work currently underway to develop broadband access standards, the time scales are incompatible with the early commencement of commercial activities. One option would be for the licence conditions to require that networks be brought into line with appropriate European standards at a later date.

A number of regulatory issues were raised with contributors to this study. Their views are summarised here.

Many, but not all, contributors were confident that the business case for Broadband Wireless Access was sound even if the carriage of POTS and broadcast services were to be prohibited. Several pointed out, however, that forcing subscribers to purchase broadband services through one operator and conventional telecommunication services through another could confuse the market. An alternative could be to prohibit Broadband Wireless Access operators from switching voice.

A prohibition on the carriage of broadcast services was not considered to be significant.

We note that the competitive broadband access operators are likely to be BT and the cable companies who are permitted to carry and switch voice. This suggests that a prohibition on the carriage of POTS would be unnecessarily restrictive. Nevertheless, it would not be unreasonable to expect that the primary business of a broadband operator should be broadband services and that there should be some obligation to this effect.

Licences could provide licensees with exclusive access to a particular block of spectrum on a national, regional or local basis. There is interest in all three with national or regional licences being the most popular. Overall, this survey suggests that regional and local licences would be of interest to the largest number of applicants. A number of contributors expressed interest in site by site licences.

The preferences expressed generally reflected the individual aspirations of the contributors. One or two argued that an operator would need a national brand identity, and therefore a national licence, is order to compete with BT and Cable and Wireless Communications. Others argued that a local operator would have strong marketing advantages.

Most contributors expected there to be at least two Broadband Wireless Access operators in any one region. It was, however, argued that a local network (covering, for example, a Cable TV franchise area) would be viable only if local licences were exclusive.

All those with an interest in becoming a network operator expected to pay for the spectrum in one way or another but the great majority were strongly opposed to auctions resulting in a single advance payment. It was argued that such auctions would make it more difficult for start ups and smaller companies to raise the necessary capital, and would limit innovation just at the point when its contribution could be greatest.

The majority of contributors were of the view that a tax on revenue (agreed or bid at the time of licensing) would be the most equitable approach. Most would be content for licence applications to be judged on the basis of the business plan, likely future revenues to the government, spectrum efficiency and innovation.

A number of potential licence bidders are likely to be small companies or start ups who have expressed concern about the way in which a simple auction process (notably involving large up front payments) could impact on their ability to gain licences. If the spectrum is auctioned, the Government will need to consider auction design issues carefully to ensure full and fair competition.

This study has found that there is a potential market for Broadband Wireless Access and that advantage of this opportunity should be taken by the Government by moving rapidly to a formal consultation with a view to issuing the first licences as soon as possible.

The following conclusions are intended to support the preparation of any consultation document and to stimulate debate on the way in which a broadband initiative might proceed. Our conclusions are drawn from the results of the interviews and assume that:

In the longer term we expect that the LDO companies will become Full Services Access Network operators, and that some will wish to enhance MVDS to provide a full range of broadband facilities. We take the view, therefore, that MVDS and Broadband Wireless Access should evolve towards a common licensing regime and, most likely, a common technical standard. The Agency should encourage this process by, for example, establishing a group of interested parties to make proposals as to how the current MVDS standard could be evolved towards a Full Services capability. The results of the CRABS and Cabsinet projects and the ETSI BRAN work will be relevant.

An illustrative suggestion as to how spectrum might be allocated in two phases in a scenario involving Pioneer, regional and local licences and taking account of the way in which the market may develop is set out below and summarised in Figure 5.1.

During the first few years, market development is likely to be confined primarily to the business sector. Initially, penetration will be low and the typical bit rates per subscriber will be of the order of 2 to 8 Mbits per second. Consequently, a total of 200 to 400 MHz of spectrum per Pioneer licence should, we estimate, be adequate.

At this stage, we suggest that operators should be expected to justify why any point-to-point link requirements they may have cannot be satisfied from within their spectrum allocation before assignments are granted in any other fixed services band.

In the slightly longer term, regional licensees are expected to roll out networks providing a full range of broadband services, both business and residential. We suggest that 1 GHz be assigned to each operator. In order to treat each equally, we suggest that this be achieved by allocating each operator spectrum from the 27.5 to 29.5 GHz and 42.5 to 43.5 GHz bands. The allocation at 43 GHz could be in the form of one or two blocks dependent on technical developments at the time.

Note that the allocation at 43 GHz would include the necessary guard bands, and that requirements for point-to-point links would be met from within the operator’s total allocation.

Local licence operators may well carry as much traffic as a regional operator and therefore should also be assigned 1 GHz of spectrum. Since geographical restrictions will be more easily accommodated with local licences, we suggest that the 40.5 to 41.5 GHz band be used. This means that local licences would not be available in those areas where an LDO licensee was making use of this band.

The above allocations would leave 2 x 448 MHz at 28 GHz and 1 GHz between 41.5 and 42.5 GHz (except where used by an LDO licensee). We believe that it would be prudent to regard this as spectrum to be held in reserve at this stage in the light of:

Figure 5.1 This figure illustrates how the spectrum might be allocated. Initially, no more than one Pioneer licence per region is issued with 2 x 112 MHz at 28 GHz. Subsequently, the Pioneer licence is upgraded to a Regional licence and a second Regional licence issued in each region. Each Regional licensee receives 2 x 224 MHz at 28 GHz plus 1 x 500 MHz at 43 MHz (dark shading). Local licensees receive 1 x 1 GHz at 41 GHz (light shading) where there is no LDO usage. Reserved spectrum is unshaded.

Bosch Telecom	Nortel Fixed Wireless Access
Cable and Wireless Communications	Norweb
Castle Transmission Int’l Ltd	Ovum Ltd
The Convergence Group plc	Philips Broadband Networks
Energis	P-Com
Ericsson Ltd	Radiant Networks plc
Eurobell plc	The Rutherford Appleton Laboratory
Formus Communications Inc	Satconsult
Granger Telecom	Siemens Roke Manor Research
Harris Systems	Silverstone Electronics Ltd
HighwayOne Corporation Ltd	Tadiran Telecommunications Ltd
Ionica International Ltd	Telegen Ltd
Libera	Timespace Radio AB
MLL Telecom Ltd	WIC Connexus Ltd
Netro Corporation	WinStar Communications Inc
Newbridge Networks Ltd	Worldpipe Ltd