This is the year of the "dark fiber" - a transparent superhighway -"an information hosepipe" unimpeded by electronic amplifiers or switches, over which gigabits-per-second of data can be delivered. Last year, the U.S. Federal Communications Commission (FCC) ruled that three of the Regional Bell Operating Companies handling local service would have to supply direct optical-fiber links between customers on an experimental basis. That ruling was reinforced by the announcement of U.S. President Clinton and Vice President Gore supporting the development of a national gigabit data highway. Meanwhile, the cost of passive optical networks fell to be on a par with that of twisted copper pairs, making fiber to the home economically competitive with traditional telephone technology for the first time.

These developments catapulted into public realization the fact that transparent optical networks will revolutionize telecommunications, computer and information networks, and even national economies. Moreover, at the same time, we are witnessing the convergence of key technologies, market needs, and costs that now permit transparent optical networks to be introduced economically. In fibre technology the key to the low cost and high bandwidth is the replacement of electronic repeaters with passive optical amplifiers. Today's electronic repeater employs a photonic detector that converts light pulses into an electronic signal that is amplified, processed, and used to drive a semiconductor laser, which generates a new optical signal for the next stage in its travel through the fiber. The limitations of the electronic circuits involved present a restricted signal bandwidth in the fibre path. In contrast, optical amplifiers are capable of boosting a wide range of wavelengths with no intervening conversion of photons to electrons. Although an optical amplifier may be essentially a semiconductor laser operated below the lasing threshold, the most revolutionary and promising type is simply a length of the optical fiber itself. By optically pumping doped or undoped fiber in a cable, amplifiers can now span 50 km, with distributed amplification offering the optimum system solution in terms of the end-to-end S/N ratio.

The introduction of optical amplifiers also allows the number of switching sites in a network to be reduced by a full order of magnitude as the signalling, transmission, and reliability constraints of their copper forebears are overcome. Economically and operationally a "fibre pull back" of switching is automatically afforded by transparent optical networks.

Among the features of a transparent and passive optical network that are not immediately obvious are the indirect cost reductions with the progressive eradication of unnecessary human activity within a network. Typically a telephone company expects to see about half of all faults in its copper local loop. About 25% are induced by re-routing, repair and maintenance action. In other words, installation and repair crews actually introduce faults directly, and also leave latent faults and other incipient damage in the course of their scheduled work. Moreover, corrosion contributes a further 10%, windage 5%, and damage by other utilities digging near an underground cable some 10%. Therefore, optical fiber, and particularly passive optical networks that do not inherently require physical re-routings as they can be re-routed by wavelength and/or time slot address. This can potentially save telephone companies some 40% of faults reports compared with copper. When this is coupled with the advantages of transparency allowing a degree of future proofing between switch, site or hub through to the home then PONs show significant economic advantage over all other solutions including fiber to the kerb and wireless to the home.

Last year also saw numerous demonstrations of microwave radio transmission directly over fiber. These revealed that radio cells can be coupled into transparent optical paths and be replicated (rebroadcast) at vast distances from the transmitter without significant intermediate technology. Even more interesting, optical wireless systems utilising semiconductor and fiber based technologies demonstrated that huge bandwidths can be realized in free space. New forms of holographic lensing (optical antennas) and optical leaky feeder have been demonstrated that allow the creation of pica cellular wireless systems from the size of a room down to the size of a pin head. With this technology, it is feasible to have office systems that are wireless, with diffuse optical fields giving a uniform signal power in a room. Alternatively, the signal may be focused on an individual seat, desk, and specific equipment. It is now possible to transmit the entire contents of the Encyclopaedia Britannica (`1Gbit of information) into a shirt pocket device in less than one second.

?

These developments are fuelling a growing contention between telephone companies and the computer network providers. Until last year, the conflict was perceived to be quite straightforward; telecommunications people talk in terms of switching, while computer users talk in terms of routers. What the computer community requires is dark fiber, or a super highway, that allows them to transmit unimpeded, unrestricted at any time without switching and electronics in the way. Conversely, the telephone companies wish to maintain control of their networks by the ownership of switches, plus the intelligence in the core of the network to control calls and the allocation of bandwidth. The reality is, neither camp is likely to win, because the requirement for a successful system design on a global scale will most likely be a meld of both.

There is no doubt that intelligence is rapidly migrating to the periphery of networks (in the form of smarter terminals and applications) and the need for such functions as switching and intelligence in the core is diminishing. This trend will likely continue and accelerate between now and the year 2000. Today, there are two key impediments to the realization and success of future data highways: first, the lack of co-operation between the telecommunications and computer industries; and second, standards that are dead on arrival, with protocols and operating systems from the computer industry that are vested in the 1960's.

Since 1980, we have seen optical fiber transmission capabilities double each year, while costs have dropped exponentially. We can expect such progress to continue beyond the year 2000. Throwing bandwidth at problems such as signalling, control and network management, naturally leads to the suggestion that wavelength division multiplexing (multiplexing signals by color of light) should supersede time division multiplexing (multiplexing by sampling signals in time). It may also be that the current contest between soliton propagation as a means of propagating undistorted signals around the planet, as opposed to wavelength division multiplexing, which incurs some significant distortion, will be resolved by a mixture of the two technologies. It may also be that ATM, or packet switching techniques, at a photonic level become the optimum for a massive global telecommunications network. Such a solution would be devoid of the delays, blocking and information waves that are inherent in ATM realised electronically over restricted bandwidths systems using radio, copper and limited bandwidth fibre. However, that suggestion now looks as revolutionary as suggesting that the code compression of signals should be abandoned because the bandwidth restrictions of copper has gone away. Who knows, in a decade we may see the realization of such radical solutions promoted by passive and transparent optical networks - and then we really will be able to telecommunicate!

Peter Cochrane (F) is Head of The Core Technologies Research Department at British Telecom Laboratories at Martlesham Heath, Ipwsich, U.K. He has a 620 strong team studying future technologies systems, networks and services. In 1990, he received the Queen's Award for Technology as manager for the production of optical receivers for TAT 8 and the PTAT 1 undersea cable systems.