Telecommunications Essentials: The Complete Global Source for Communications Fundamentals, Data Networking and the Internet, and Next-Generation Networks

Fiber Optics

In the late 1950s and early 1960s, a number of people were working in the realm of fiber optics simultaneously. Charles Kao, who was a scientist with ITT, is often acknowledged as being one of the fathers of fiber optics. Kao theorized that if we could develop a procedure for manufacturing ultrapure, ultrathin filaments of glass, we could use this as a revolutionary new communications pipeline. Thus began the move toward researching and developing optical technology.

In 1970, we first had the developments that today allow us to deploy the large amounts of fiber we have. The first development was a procedure for manufacturing ultrapure filaments of glass, a procedure called "broomsticking," that was introduced by Corning Glassworks. Glass that has an inner core etched into it is melted at extremely high temperatures, and as the glass is melting and dropping down the tube, it begins to cool and form a strand. By the time it gets to the bottom of the tube, it is a fiber-optic thread. Being able to create the fiber cable itself solved half the equation, but because the fiber's diameter is minuscule (measured in micrometers, or microns, abbreviated ), the light source that will pulse energy on this tiny fiber also has to be minuscule. In 1970, Bell Labs completed the equation by introducing the first laser diode small enough to fit through the eye of a needle. So the two things that determine the performance characteristics of a given fiber implementation are the type of cable used and the type of light source used.

Characteristics of Fiber Optics

Fiber optics operates in the visible light spectrum, in the range from 10¹⁴Hz to 10¹⁵Hz. Wavelength is a measure of the width of the waves being transmitted. Different fiber-optic materials are optimized for different wavelengths. The EIA/TIA standards currently support three wavelengths for fiber-optic transmission: 850, 1,300, and 1,550 nanometers (nm). Each of these bands is about 200nm wide and offers about 25THz of capacity, which means there is a total of some 75THz of capacity on a fiber cable. The bandwidth of fiber is also determined by the number of wavelengths that it can carry, as well as by the number of bits per second that each wavelength supports. (As discussed in Chapter 2, each year, wavelength division multiplexers are enabling us to derive twice as many wavelengths as the year before, and hence they enable us to exploit the underlying capacity of the fiber cables.)

With fiber, today we can space repeaters about 500 miles (800 kilometers) apart, but new developments promise that, in the very near future, new techniques will stretch that spacing. Trials have been successfully completed at distances of 2,500 miles (4,000 kilometers) by Xros (now Nortel) and, more recently, at a distance of 4,000 miles (6,400 kilometers) by Williams Communications (using Corvis).

Components of Fiber Optics

As mentioned earlier, the two things that determine the performance characteristics of a given fiber implementation are the type of cable used and the type of light source used. Let's look at the components of each.

Fiber-Optic Cable

Figure 3.9 shows the basic components of fiber-optic cable. Fiber-optic cable is available in many sizes. It can have as few as a couple pairs of fiber or it can have bundles that contain upward of 400 or 500 fiber pairs. Each of the fibers is protected with cladding, which ensures that the light energy remains within the fiber rather than bouncing out into the exterior. The cladding is surrounded by plastic shielding, which, among others things, ensures that you can't bend the fiber to the point at which it would break; the plastic shielding therefore limits how much stress you can put on a given fiber. That plastic shielding is then further reinforced with Kevlar reinforcing material very strong material that is five times stronger than steel to prevent other intrusions. Outer jackets cover the Kevlar reinforcing material, and the number and type of outer jackets depend on the environment where the cable is meant to be deployed (for example, buried underground, used in the ocean, strung through the air).

Figure 3.9. Fiber-optic cable

There are two major categories of fiber: multimode and monomode (also known as single mode). Fiber size is a measure of the core diameter and cladding (outside) diameter. It is expressed in the format xx/zz, where xx is the core diameter and zz is the outside diameter of the cladding. For example, a 62.5/125-micron fiber has a core diameter of 62.5 microns and a total diameter of 125 microns. The 62.5/125-micron fiber is the only size currently supported by the EIA/TIA standard.

The core diameter of the fiber in multimode ranges from 50 microns to 62.5 microns, which is large relative to the wavelength of the light passing through it; as a result, multimode fiber suffers from modal dispersion (that is, the tendency of light to travel in a wave-like motion rather than in a straight line), and repeaters need to be spaced fairly close together (about 10 to 40 miles [16 to 64 kilometers] apart). The diameter of multimode also has a benefit: It makes the fiber more tolerant of errors related to fitting the fiber to transmitter or receiver attachments, so termination of multimode is rather easy.

The more high-performance mode of fiber, single-mode fiber, has a fiber diameter that is almost the same as the wavelength of light passing through it from 8 microns to 12 microns. Therefore, the light can use only one path it must travel straight down the center of the fiber. As a result, single-mode fiber does not suffer from modal dispersions and it maintains very good signal quality over longer distances. Therefore, with single-mode fiber, repeaters can be spaced farther apart (as mentioned earlier, they are currently about 500 miles [804 kilometers] apart, with the distances getting greater rapidly). But because single-mode fiber has such a smaller diameter, it is difficult to terminate, so experienced technical support may be needed to perform splices and other work with single-mode fiber.

The bottom line is that multimode fiber is less expensive than single-mode fiber but offers lower performance than single-mode fiber. Single-mode fiber is more expensive and offers higher performance, and it has been used in most of the long-distance networks that use fiber.

Light Sources

In the realm of light sources, there are also two categories: light-emitting diodes (LEDs) and laser diodes. The cheaper, lower-performer category is LEDs. LEDs are relatively inexpensive, they have a long life, and they are rather tolerant of extreme temperatures. However, they couple only about 3% of light into the fiber, so their data rates are low, currently about 500Mbps.

Laser diodes are capable of much higher transmission speeds than LEDs. They are a pure light source that provides coherent energy that has little distortion. Laser diodes, therefore, are commonly used for long-haul and high-speed transmission. Laser diodes offer better performance than LEDs, and they are more expensive, although the cost of these components has been dropping about 40% per year. As the costs drop, performance is also improving; in the very near future, we should see the introduction of light sources that pulse one trillion bits per second.

So, when you want to carry traffic over the long haul, the best combination is single-mode fiber with laser diodes. For very short implementations, such as in a campus network environment, the cost-efficiencies of multimode and LEDs may make this combination a more appropriate solution. But in general, as we look forward to the new developments in optical equipment such as wavelength division multiplexers, optical cross-connects, and optical switches we will need higher-quality fiber to interface to. It appears that roughly 95% of the world's fiber plant is not prepared to operate at the high speed that we are evolving to with optical equipment. So, even though we have been actively deploying fiber for years, not all of it is compatible with the next generation of optical equipment. This means that we will see new companies laying new highways and using the latest and greatest in fiber, as well as older companies having to upgrade their plants if they want to take advantage of what the optical equipment has to offer.

How Fiber-Optic Transmission Works

As shown in Figure 3.10, in fiber-optic transmission, the digital bit stream enters the light source, in this case the laser diode. If a one bit is present, the light source pulses light in that time slot, but if there is a zero bit, there is no light pulse. The absence or presence of light therefore represents the discrete ones and zeros. Light energy, like other forms of energy, attenuates as it moves over a distance, so it has to run though an amplification or repeating process. As mentioned earlier, until about 1994, electronic repeaters were used with fiber, so the optical signal would have to stop; be converted into electrical energy; be resynchronized, retimed, and regenerated; and then be converted back into optical energy to be passed to the next point in the network. This was a major problem because it limited the data rate to 2.5Gbps. But some developments were introduced in the early 1990s that dramatically changed long-distance communications over fiber. The next section of this chapter talks about these innovations.

Figure 3.10. Fiber-optic transmission

Innovations in Fiber Optics: EDFAs, WDM, and DWDM

As mentioned in Chapter 2, erbium-doped fiber amplifiers (EDFAs) are optical repeaters that are made of fiber doped with erbium metal at periodic intervals (normally every 30 to 60 miles [50 to 100 kilometers]). The introduction of EDFAs made it possible for fiber-optic systems to operate at 10Gbps. EDFAs also opened the way for Wavelength Division Multiplexing (WDM), the process of dividing up the optical transmission spectrum into a number of nonoverlapping wavelengths, with each wavelength supporting a single high-speed communications channel. Today, undersea cables need to be designed with WDM in mind, and until recently, most were not, which means they have inappropriate repeater spacing. So, again, for the next generation of fiber communications over undersea cables, many systems will have to be replaced or upgraded.

An optical multiplexing hierarchy was the predecessor to WDM: Synchronous Digital Hierarchy (SDH) and Synchronous Optical Network (SONET). SDH/SONET is a time-division multiplexed system, and within SDH/SONET fiber cables, you can carry energy over just one wavelength. (SDH/SONET is discussed in detail in Chapter 5, "The PSTN.") With Dense WDM (DWDM), you can operate over 16 or more wavelengths. Products that are currently shipping support anywhere from 80 wavelengths to 128 wavelengths and operate at data rates of 2.5Gbps to 10Gbps. New systems are emerging that operate at 40Gbps. Meanwhile, research is also under way with dense wavelength division multiplexers that will be capable of supporting as many as 15,000 wavelengths. These developments are just the tip of the iceberg with what we can expect in coming years. (WDM and DWDM are discussed in more detail in Chapter 12, "Optical Networking.")

Applications of Fiber Optics

Fiber has a number of key applications. It is used in both public and private network backbones, so the vast majority of the backbones of the PSTNs worldwide have been upgraded to fiber. The backbones of the Internet providers are fiber. Cable TV systems and power utilities have reengineered and upgraded their backbones as well.

Surprisingly, electric power utilities are the second largest network operator after the telcos. They have vast infrastructures for generating and transmitting electricity; these infrastructures rely on fiber-optic communications systems to direct and control power distribution. After they have put in fiber, they have often found themselves with excess capacity and in a position to resell dark fiber to interested parties. When you lease dark fiber, you're basically leasing a pair of fibers, but you don't have the active electronics and photonics included with it, so you are responsible for acquiring that equipment and adding it to the network. But with dark fiber, you're not paying for bandwidth you're paying for the physical facility and if you want to upgrade your systems to laser diodes that pulse more bits per second, or if you want to add a wavelength division multiplexer to access more wavelengths, these changes will not affect your monthly cost for the fiber pair itself. Power utilities have been big players in the deployment of fiber throughout the world.

Another application of fiber is in the local loop. There are numerous arrangements of fiber in the local loop, including HFC (that is, fiber to a neighborhood node and then coax on to the subscribers); fiber to the curb with a twisted-pair solution to the home; fiber to the home that terminates on its own individual optical unit; and passive optical networking, which promises to greatly reduce the cost of bringing fiber to the home. Chapter 15 covers the details of these various arrangements.

Another application for fiber is in LANs. Fiber Distributed Data Interface (FDDI) was the first optical LAN backbone that offered 100Mbps backbone capacity, but today it has largely been displaced by the use of 100Mbps and Gigabit Ethernet, both of which can be accommodated over fiber.

Another application of fiber involves the use of imagery or video when extremely high resolution is critical (for example, in telemedicine). Consider an application that involves the transmission of images between an imaging center and a doctor's office. Say you went to the imaging center to have an x-ray of your lungs, and in the transmission of your lung x-ray, a little bit of noise in the network put a big black spot on your lung. If that happened, you would likely be scheduled for radical surgery. So, for this type of application, you want a network which ensures that very little noise can affect the resolution and hence the outcome of the analysis. A lot of this use of fiber occurs in early-adopter environments where there are applications for imaging, such as universities, health care environments, and entertainment environments.

Another frontier where fiber is now being investigated is in home area networks (HANs). This is a very interesting area because when broadband access comes into your home, you see a shift in where the bottleneck resides, from the local loop to inside the home. Broadband access into the home requires a broadband network within the home to properly distribute the entertainment, data, and voice services that are collectively being transported over that broadband access into the home. So, there are ways to wire a new home with fiber or to retrofit an older home so that you can enjoy really high-quality entertainment and data networks within the home. (HANs are discussed in more detail in Chapter 15, "The Broadband Home and HANs.")

Advantages and Disadvantages of Fiber Optics

The advantages of fiber optics are as follows:

         Extremely high bandwidth Fiber offers far more bandwidth than any other cable-based medium.

         Elastic traffic-carrying capacity Without having to change the fiber, assuming that it's the correct generation of fiber, you can add equipment that provides additional capacity over the original fiber alone. This, along with DWDM's capability to turn various wavelengths on and off at will, enables dynamic network bandwidth provisioning to accommodate fluctuations in traffic.

         Not susceptible to electromagnetic impairments or interference Because fiber is not susceptible to electromagnetic impairments or interference, it has a very low bit error rate, 10 ¹³, which means fiber-optic transmissions are virtually noise free.

         Secure transmission and early detection By constantly monitoring the optical network, you can detect when the light pulse diminishes, enabling you to detect leaks.

         Low in weight and mass Because fiber is low in weight and mass, much less human installation power is needed than with traditional copper cable or coax bundles.

The disadvantages of fiber optics include the following:

         High installation costs, but dropping Fiber installation is still relatively costly, although the cost has been dropping by about 60% per year, depending on the components. So although the price keeps getting better, it can still mean a fairly impressive investment. For example, the capital cost of bringing fiber to the home, including the construction efforts, is about US$2,500 per subscriber. But it is also anticipated to be down to about US$1,200 to US$1,500 per subscriber by around 2003.

         Special test equipment required When you start putting in fiber, you have to acquire specialized test equipment because none of the test equipment you use on an electrical network will work with fiber. You need an optical time-domain reflectometer (OTDR), and when you get into more sophisticated optical networks, you need highly specialized optical probes that can cost US$1 million and you need one at each location.

         Shortage of components and manufacturing sites We currently have a shortage of components and a shortage of manufacturing sites to make more fiber. Those that don't already have contracts negotiated for the delivery of fiber may have a very hard time getting more over the next year, until we ramp up the manufacturing facilities to take on the growing demand. This is also the case with the production of optical components, such as optical switches and optical wavelength routers.

         Vulnerability to physical damage Fiber is a small medium, so it can very easily be cut or otherwise damaged (for example, in a railroad car derailment, in the midst of construction activities). When you choose fiber as the primary medium, you have to address backup, restoration, and survivability from the start because the likelihood of damage is great.

         Vulnerability to damage caused by wildlife A number of flora and fauna cause damage to fiber. A number of birds really like the Kevlar reinforcing material and think it makes lovely nests for their babies, so they peck away at fiber-optic cables to get at that Kevlar material. At least five different types of ants seem to enjoy the plastic shielding in their diet, so they nibble at the underground fibers. Sharks have been known to chomp on cable near the repeating points. A plant called the Christmas tree plant thinks that fiber-optic cable is a tree root and wraps itself around it very tightly and chokes it off.

Wiring for Tomorrow: Undersea Fiber Cables

The first undersea cable, which was laid for telegraph, was laid in 1851 between England and France. In 1956 the first coax cable called transatlantic link (TAT-1) went in. TAT-1 had the capacity to carry 35 conversations over 64Kbps channels.

The first fiber undersea cable, laid in 1988, was called TAT-8 and could support 4,000 voice channels. But undersea use of fiber didn't take off until 1994, when optical amplifiers were introduced. By the end of 1998, some 23 million miles of fiber-optic cable had been laid throughout the world, by dozens of companies, at tremendous cost. By mid-1999 the total transatlantic bandwidth was 3Tbps, compared to just 100Gbps in 1998. By the end of 2001, we are likely to reach 6Tbps. Between Asia and Europe, in 1997, we had bandwidth of 11Gbps; by 1999, we had 21Gbps; and by 2003, we should have 321Gbps. You can see that a great deal of spending has been done on fiber-optic cable, all over the world.

Fiber technology breakthroughs are having a profound impact on service providers, and that's witnessed by the constantly changing prices for intercontinental capacity. The construction cost of 64Kbps circuits has dropped from almost US$1,500 in 1988, to US$300 in 1995, to just a couple dollars per line today. When operators purchase undersea capacity, they pay two charges. The first is a one-time charge for the bandwidth the indefeasible right of use. The second is an ongoing operations, administration, and maintenance charge that's recurring for the maintenance vessels that service the cable, and this is typically 3% to 5% of the total purchase cost.

The economic shifts look like this for a capacity of 155Mbps: At the start of 1997, it would have cost US$20 million; in 1998, it was down to US$10 million; in early 2000, it was down to US$2 to US$3 million; and in 2001, it's expected to be at US$1 million. The operations, administration, and maintenance charges have remained the same because the contracts originally called for the calculation of those charges based on the cable length as well as the bandwidth, so as you increased your bandwidth, your operations, administration, and maintenance charges increased. Those agreements were recently changed so that you are only charged the operations, administration, and maintenance costs for the length of the cable. Hence, as you expand capacity, the maintenance charge is dropped.

As you can see, there's a great deal of excitement about wiring the world with fiber in preparation for accommodating tomorrow's advanced applications. Chapter 12 elaborates on the range of equipment that's being introduced to allow us to take advantage of the vast potential of the fiber-optic spectrum.