Using light as a form of communication is by no means a recent development. One of the first forms of optical communication was using fire and smoke signals. Technically speaking, a simple gesture such as waving your hand comes under the purview of optical communication since it requires the presence of light - performing the same gesture in a dark room will not evoke a response from the other party. The modern descendants of fire and smoke signals are lamps and other advanced light sources for signalling applications such as lighthouses for ships, and beacons and night-time runway guidance lights for aircraft. The major problems with the aforementioned lightwave systems are that the rate of information transfer is very slow and a direct line-of-sight is required between the source and the recipient.

A fiber optic system for long-distance communication principally consists of the following: an optical transmitter (typically a semiconductor laser); optical fiber made of silica-based glass - the fiber acts as the medium by which the information is transferred; optical amplifiers e.g., erbium-doped fiber amplifiers (EDFAs) and/or Raman amplifiers; an optical receiver such as a photodetector. Figure 1 is a schematic showing the principal components of a simple fiber optic communication system. There are also a gamut of other components in the system such as modulators, switches and multiple-xers/demultiplexers that play crucial roles in system operation and performance. We will briefly look at each of the principal components of the fiber system.

The information transfer process in fiber optic systems begins at the optical transmitter that is typically a laser source. The laser converts the information that is originally in the form of an electrical signal to an optical signal. The optical signal then travels through the fiber to the desired destination. The optical fiber is a "pipe" for transporting light. Light in a fiber is essentially trapped within it because the core of the fiber has a higher refractive index than the cladding as shown in Figure 2. As the optical signal traverses the fiber, it diminishes in power because of loss in the fiber. Optical amplifiers are then used periodically every few tens of kilometers to boost the power of the optical signal to compensate for the loss. Finally at the destination, the photodetector (optical receiver) converts the received optical signal back to an electrical signal. The concept is straightforward and elegant; however, its implementation is by no means trivial but rather complicated. We will now focus on some important fiber characteristics that dictate the performance requirements for other devices in the system.

An important design issue for communication systems is its loss. A very lossy system is not desirable because the signal will require frequent amplification before it reaches its destination (optical amplifiers are expensive!). The loss in a fiber depends on the wavelength (or frequency) of the light used to transmit information in the system. For purposes of illustration, imagine that if you used blue instead of red light to transmit your information, you would incur a higher loss in the fiber. Real fiber systems actually do not use light with wavelengths in the visible spectrum (the visible spectrum is light that the human eye can detect such as blue, green and red) for a variety of reasons, one of which is that these range of wavelengths exhibit a high loss in the fiber. The lowest loss in a silica-based fiber is achieved in the infrared at a wavelength of approximately 1550 nm (one billion nanometers makes a meter). The fiber loss at this wavelength is about 0.2 dB/km (this corresponds to a loss of about 4.5 % of the signal power every kilometer). Wavelengths around 1550 nm are today the standard for long-haul (greater than a few hundred kilometers) fiber systems. It is worth mentioning that for smaller networks such as local, campus and metropolitan, the wavelengths 1310 nm and 850 nm are more popular. You may ask, why would one want to operate at wavelengths with a higher loss? Loss is not the major concern for smaller networks because the total fiber loss is low owing to the shorter distances involved. Apart from loss, there are other technical and economic factors that need to be taken into consideration when deciding on the wavelength of operation. Without delving into too much detail, 1310 nm has the advantage of exhibiting a lower dispersion which implies reduced pulse spreading, and 850 nm has the advantage that it is easier and more economical to produce laser sources in this regime.

Optical fibers have distinct advantages over their metallic counterparts, copper cables (both twisted pair and coaxial). Optical fibers are compact, lightweight and have the ability to transfer large amounts of information. This directly leads to economical advantages. One should note that the economic figure of merit for installing communication systems is not the cost per unit distance ($/meter) but rather the cost per unit distance per unit information transfer ($/meter/bit). To a first-order approximation, optical fibers are 200 times lighter, 150 times less in volume and 10,000 times greater in bandwidth than coaxial cables. With these numbers in mind, the overwhelming advantages of optical fiber technology become clearly apparent. There are, however, a few challenges that optical fibers have to deal with compared to copper cables working with fibers requires a great deal of skill, care and time. Despite how surprising it may sound, one of the reasons you will not see fiber-to-the-home (FTTH) become a reality soon is that the fiber connectors that need to be attached in every home is not an automated process but rather a labour-intensive one that requires specialized personnel.

Apart from the economic advantages, optical fibers also possess technological superiority. Optical fibers, be they silica-based glass or plastic, are insulators and, therefore, have no currents flowing in them. As a result, fibers are immune to electromagnetic interference. Sources of electromagnetic interference include, but are not limited to, cellular base stations, radio and television stations, radar, high-voltage power lines, lightning and nuclear explosions. In addition, fiber systems cannot be tapped into without being detected. This degree of security is essential for systems transferring sensitive information such as national security and military issues. Another advantage of optical fibers is that they do not corrode.

The demand for optical fiber today is almost unimaginable. Fiber is being deployed worldwide at a rate of about 5,000 km per hour! The giant fiber manufacturers such as Corning and Lucent Technologies are sold out of their fiber for the next 11 to 12 months. Data and voice traffic are increasing at a rate of about 80 per cent and 10 per cent per year, respectively. The extensive research on optical fiber technology and its supporting electronics has resulted in our ability to double the data-carrying capacity of each fiber. The aggregate bit rate of state-of-the-art commercial fiber systems now available allow a data transfer capability of 400 Gb/s (one gigabit is equal to one billion bits) on a single fiber; for the non-technical, 400 Gb/s translates to transporting about 12,000 encyclopaedia volumes per second. Experiments have been demonstrated in laboratories where the aggregate bit rate was 1.6 Tb/s (one terabit is equal to one trillion bits) over a few hundred kilometers.

The aggregate bit rate of fiber systems is currently so high because of two principal reasons: the bit rate per wavelength and number of wavelengths per fiber. On a certain wavelength (or colour if you will), the bit rate is determined by the number of times you can turn your light source on and off per second. For technical reasons, the laser source in high-speed systems is actually kept constantly on and the on-off function is performed by a subsequent device, a modulator that acts as a shutter. Current technology is at a bit rate of 40 Gb/s. One of the reasons it will be difficult to go beyond this speed is that the present supporting electronics technology cannot operate at speeds beyond 40 Gb/s with the required performance criteria. This problem is a rather new phenomena because in the past, optical technology was always behind electronics. Higher bit rates will be achieved with advancements in electronics and optics.

Imagine that you have a fiber system using red light to transmit your information at 40 Gb/s (as mentioned earlier, using red light is just for illustrative purposes and has nothing to do with reality). If you could now simultaneously send blue light (a different wavelength) at 40 Gb/s with the red light, your aggregate bit rate would be twice as much as before. The more colours or wavelength you can simultaneously transmit, your aggregate bit rate increases proportionally. The technology of simultaneously transmitting multiple wavelengths is used in modern fiber systems and is known as wavelength division multiplexing (WDM). WDM systems with varying number of wavelength e.g., 16, 40 and 80 have been deployed or experimentally tested at different bit rates. The experiment earlier mentioned with an aggregate bit rate of 1.6 Tb/s involved transmitting 40 wavelengths each at a speed of 40 Gb/s. The proximity and number of wavelengths (wavelength density) that can be used in the allowed practical frequency range is determined and limited by a number of factors including pulse broadening owing to dispersion and nonlinear effects such as four-wave mixing and self-phase modulation. Researchers continue to explore new avenues to overcome the challenges presented by dispersion and nonlinearities.

There have been great strides in progress of fiber optic technology, especially since the early 1970s. The focus of research and development has evolved from reducing the losses in a fiber to trying to perform all-optical networking, the holy grail of fiber optic communication. Fiber optics has found applications in telephone trunks, subscriber service, broadcast and cable TV, data communication and sensors. The growth of the internet has fuelled the demand for bandwidth. One approach to meeting the demand for bandwidth has been installing more fiber, albeit not the most economical one. Another approach has been to increase the bandwidth of existing fibers. As optical fiber communication researchers and scientists continue to push the envelope of the capabilities of this technology, we will continue to experience the effects of the communications revolution. Researchers at Bell Laboratories project that there will be enough installed bandwidth by the end of this decade to provide every person on the planet 0.1 Gb/s continuously. Where this technological revolution will take us and the extent of its socio-economic impact is almost anybody's guess. One thing, however, is certain the world will surely be a smaller place to live in.