No strings

Fixed wireless is standing in the wings, ready to relieve broadband service providers of their dependence on incumbent carriers,...

Fixed wireless is standing in the wings, ready to relieve broadband service providers of their dependence on incumbent carriers, but will it trump DSL?

Fixed wireless could provide the most cost effective way of accessing networks at broadband speed, despite the technology's mixed success so far. The application of fixed wireless links for data communication has until now mostly been confined to low speeds and to locations beyond the reach of alternative wire-based options.

But emerging fixed wireless technologies promise to compete more strongly with the other contenders, such as digital subscriber line (DSL) and cable modems, while providing more extended coverage and freedom from the incumbent owners of telephony or cable TV networks.

This could usher in an era of genuine choice and competition for broadband data services between a number of providers each in control of its own network on an end-to-end basis without dependence on incumbent carriers for the final local loop. For users, this brings the potential of more varied and lower cost services.

But there are some technical and regulatory hurdles to be overcome before fixed wireless becomes a serious contender for broadband access. At first sight the omens are not good, certainly to judge from the story so far.

The first wireless services in the mid-1990s were all narrowband and focused mainly on voice. Many were unsuccessful, partly because of problems of capacity and reliability. One of the UK's largest fixed wireless companies, Ionica, collapsed spectacularly. There was limited mileage in such narrowband services for data, in the business sector and attention switched to broadband.

It appeared necessary to move up to higher frequencies to achieve broadband data rates, which are generally defined as being 512kbps or more per user. However, this reduces the range to about 3km and it also decreases the proportion of locations within that range that are capable of being reached to between 20% and 50% of all buildings, depending on terrain. The remainder will lie outside the line of sight.

Therefore, to achieve widespread coverage, many masts have to be deployed, with obvious cost and environmental implications. Nevertheless, companies and governments have proceeded with development of high-frequency fixed wireless technologies at 10GHz, 28GHz and 40GHz, culminating in the UK with the auctioning of 28GHz licences at the end of 2000. This was overshadowed by the third-generation (3G) mobile auctions earlier in the year, and went almost unnoticed among the general media.

The 28GHz spectrum auction was a bigger flop than the 3G mobile auction was a success - it raised just £38.1m against the government's expectation of £1bn. This was partly a reaction to the dotcom and high tech collapse that was by then well under way.

The major factor, however, was a growing realisation that the technology, called local multipoint distribution service (LMDS), would fail, says Peter Wharton, chief executive of fixed radio system maker Cambridge Broadband.

For this reason, Cambridge Broadband is focusing on an alternative technology at the lower frequency of 3.5GHz, which achieves greater range and coverage without sacrificing performance, according to Wharton. The evidence so far is that he is correct.

There are five criteria for judging the success of a fixed wireless technology: range; percentage of buildings that can be hit by the signal by virtue of being sufficiently within line of sight; data rate; cost; and quality of service. The 3.5GHz technology now being trialled in the Cambridge area has a good aggregate score on all five counts.

"We can do twice the speed of LMDS, match what it can do on quality of service, but with the cost advantages of lower frequencies, and a range of 20km instead of 3km, along with 90% coverage of buildings instead of 20% to 50%," says Wharton.

Furthermore, Cambridge Broadband has elegantly solved the backhaul problem common not just to all fixed wireless services, but to DSL as well. The issue here is that having provided access over the last few kilometres beyond the destination, the service provider then has to connect back into its own, probably fibre-based, trunk network, at sufficient capacity.

For DSL providers, the only recourse in most cases is to lease a circuit from the incumbent,which in the UK is BT. The same applies for fixed wireless providers, but Cambridge Broadband does the backhaul by radio transmission within the same allocated 3.5GHz frequency spectrum, using the same equipment. This "self-backhaul" approach eliminates dependence on any incumbent carrier, with potential benefits for competition and user price.

But now for the bad news. All these potential advantages are being jeopardised by the dragging of regulatory feet and failure to provide true competition, says Wharton. Auctions for potential service providers in the 3.5GHz spectrum, as opposed to the 28GHz range allocated in December 2000, had been promised for this year but have been repeatedly delayed. It also appears that the Government favours having just one national provider in order to concentrate the customer base on just one company to reduce the risk of further Ionica-type failures.

However, studies have shown that in any given segment, such as mobile telephony, at least four alternative competing service providers are needed to avoid planned or unplanned cartels forming and to ensure that there is real price-cutting competition. Of course, there will be competition from DSL, but even then there are likely to be only two or three genuine alternatives for any given region, says Wharton.

Indeed, he is exasperated by the UK regulators. "It is a disgrace," he says. "They are the last of the Stalinists, favouring cosy duopolies rather than genuine competition that benefits the customer."

There is also the issue of persuading customers that fixed radio is a real contender, especially when DSL is also available. Both are ultimately intermediate technologies, with the end game being direct fibre connection, which brings the greatest potential performance and reliability. But currently only 5% of UK businesses are within reach of fibre, and it will be many years before coverage becomes widespread.

The UK is already populated with swathes of fibre trunks, but businesses have to lie within corridors of about 100m either side of these, usually in major cities, for direct access to be economical. The role of DSL and fixed wireless is to radiate this bandwidth out either side of these corridors.

Until the new 3.5GHz services become available in a year or two, fixed wireless will never compete on price with DSL, says Matthew Hatton, an analyst with CIT Information & Analysis, which recently produced a report on fixed wireless. "It is considerably more expensive, so if DSL is available, I would go for that." So at present, fixed wireless is only a contender where DSL is unavailable, but the new 3.5GHz version is definitely worth watching.

Comparison of fixed access technologies
  • Optical fibre

  • Pros: Highest bandwidth, up to 10gbps. Greatest reliability

    Cons: Most expensive, and only available in large cities, confining it to big sites of major enterprises at present

  • Satellite

  • Pros: Offers similar data rates to DSL, can reach remote areas, and is cost effective for multicast applications delivered to multiple users

    Cons: Not yet widely available, and very expensive for two-way communications

  • Asymmetric digital subscriber line (ADSL)

  • Pros: Provides sufficient speeds, typically 2mbps upstream and 256kbps downstream, for most multimedia applications, is far cheaper than leased line services, whether cable or fibre based, and is potentially available in most places where there is a telephone line

    Cons: Roll out in UK slow and coverage patchy. For competition, relies on unbundling of the local loop, which has not been an unmitigated success anywhere in the world

  • Cable modems

  • Pros: More widely available in some cities than ADSL, with comparable capacity. Relatively cheap

    Cons: Mostly for residential users, with relatively few major business sites within reach. Shared bandwidth means performance degrades as more users are connected to a given local segment

  • Broadband wireless

  • Pros: Comparable bandwidth to ADSL with potentially wider reach. New 3.5GHz technology solves many of the problems, such as poor hit rate within the supposed distance range. Faster than DSL to roll out, easier to connect up new users, and allows alternative carriers to compete with end-to-end service, bypassing BT's network

    Cons: Existing 28GHz technology has poor coverage, while new 3.5GHz has yet to be approved. Competition within the new 3.5GHz spectrum may be limited.

Read more on IT legislation and regulation