Shining Light on Optical Networking

With demand for network bandwidth exploding and 10 Gigabit Ethernet going all fiber all the time, the future of optical networking looks bright. If you need stellar knowledge, check out

August 19, 2002

10 Min Read
NetworkComputing logo in a gray background | NetworkComputing

Single-mode fiber, on the other hand, can traverse hundreds or even thousands of kilometers, depending on the light source, wavelength and amount of data to transmit. Visually, telling the difference between multimode and single-mode fiber is nearly impossible, so most manufacturers make it easy on us. Multimode fiber is typically clad in an orange sheath, while single-mode is covered in yellow material.

The Light

There are two possible light sources for optical transmission: LED or laser. LEDs are used only in multimode systems, while lasers are the source of choice for single-mode fiber. LEDs cannot operate over single-mode fiber as the light is not coherent enough to travel down such a small passageway--imagine trying to fill an eye-dropper-size opening with water being poured from a bucket.

Data travels down copper lines as a series of on and off pulses to specify whether a bit is a 1 or a 0. Fiber works in the same way, but by pulsing the light source on and off. By pulsing the light faster, more data can be sent down the fiber. There is a limit to how quickly a laser can be turned on and off and still generate enough photons of a required strength to reach the other end and be interpreted. For this reason, high-speed fiber connections rely on multiple wavelengths to transmit data, combining the best of two worlds.

Multiple data transmissions can travel over copper by chopping the available bandwidth into channels. Each channel is then allowed to send data for a certain amount of time, transmitting in a round-robin fashion and giving everyone an equal chance. This method is called TDM (Time Division Multiplexing). Another transmission method is FDM (Frequency Division Multiplexing), which allows multiple transmissions by requiring each one to operate within a given frequency range. TV and radio signals are examples of FDM.

Most fiber systems combine these two methods, letting multiple frequencies carry multiple channels of data. Tunable laser diodes are used to create this WDM (Wavelength Division Multiplexing) combination. WDM is just one of three classifications for combining multiple wavelengths, or lambdas, onto a single fiber. WDM actually sits in the middle and defines the means of combining frequencies that are 10 nm apart from one another. DWDM (Dense Wavelength Division Multiplexing) is the far end of the spectrum, defining wave separation that is no more than 1 nm, with some systems running as close as 0.1 nm--a hundredfold increase over WDM. Because of this tightness in spectrum, systems using DWDM tend to be very expensive.

At the opposite end of the cost spectrum are new systems that are being designed around CWDM (Coarse Wavelength Division Multiplexing). The individual light frequencies are at least 20 nm apart, with some spaced as far as 35 nm apart. CWDM systems are being used in the metro area and even on LANs, where cost is a factor. CWDM is also not constrained to any one portion of the spectrum and operates between 1300 nm and 1600 nm, while DWDM systems usually operate above 1500 nm.

Going the Distance

Getting light pulses from Point A to Point B is the purpose of the exercise, but doing so involves some creativity. While light in fiber travels at about 200,000 kilometers per second, no light source can actually travel that far and still be interpreted as individual 1s and 0s. This is not just because of dispersion; photons also can be absorbed by the cladding that is used to surround the fiber core, decreasing the number that reach the receiver. Because the receiver counts the photons that reach it in a given period, with anything above a certain number considered a 1 and anything below considered a 0, fewer photons could result in false 0s.To get all the photons to the receiving end, you might think you'd raise the input power of the laser, thereby increasing the number of photons that reach the other end. This would be true if fiber optics worked in a linear fashion, but they are nonlinear. To keep input power and output power as close to linear as possible, most single-mode lasers are kept at about 0.5 watts. Increasing the power can decrease the output power because of electromagnetic forces in the core.

Integrated Bragg GratingClick here to enlarge

Regeneration is therefore the common method of extending the reach of the photons in fiber. Depending on the wavelengths used, regenerating an optical signal can take two forms: OEO (optical/electrical/optical) or FA (fiber amplifiers). OEO systems, also called optical repeaters, take the optical signal, demultiplex it and convert it to electrical pulses. The electrical signal is amplified, groomed to remove noise and converted back to optical. It is then multiplexed back on the line and sent on its way. OEO can be used across the spectrum and is commonly used for CWDM and some WDM transmissions.

Fiber amplifiers offer a more elegant solution without converting the photons to an electrical signal. The most common amplifier is the EDFA (erbium-doped fiber amplifier). Although fiber amplifiers commonly use erbium, other elements, such as praseodymin and ytterbium, can provide the same results. A better term would be REDFA (rare-earth-doped fiber amplifier), which operates by doping a section of fiber with a rare earth element, such as erbium. Doping is the process of adding impurities during manufacture; fiber-optic cable already has almost 10 percent germanium oxide as a dopant to increase the reflective index of the silica glass.

The doped section of fiber combines the signal from the transmitter with that coming from a pump laser. For erbium, the pump laser operates at 980 nm. The pump laser excites the erbium atoms, and when they are struck by photons from the original signal, some energy is transferred to the transmitted signal, amplifying it. A typical REDFA has a coil of 10 meters of doped fiber for amplification. The pump laser can be located locally to the doped fiber or remotely, as long as the strength of the light has not degraded too badly.REDFAs are typically used for submarine cables because they require much less additional equipment. REDFAs also do not add any appreciable delay to the signal, as do OEO repeaters. REDFAs, however, amplify noise in the photonic signal, and over a long distance this noise can hurt the signal.

While the problem with optical transmissions is usually generating enough strength to get to the destination, sometimes the opposite is true. When too many photons get to the receiver, it is blinded--as though it's looking directly into the sun. Adding more photons is not a good thing for receivers. Multimode receivers are immune, for the most part, to being blinded, but single-mode receivers are not. This effect typically happens when the transmitter is mismatched to the distance that needs to be covered.

Single-mode transmitters come in a variety of strengths to transmit over different distances. The maximum distance varies by bandwidth and manufacturer. Although a transmitter and receiver might be designed for the long haul, it would be disastrous to assume that they could operate at shorter distances without the receiver burning out. The receiver design assumes a certain amount of photonic loss over the distance for which the transmitter is designed. When distance or power is mismatched you need a pair of fiber-optic sunglasses, or optical attenuators. A small sliver of shaded glass is placed in the path between two sections of fiber or at the receiver. The attenuator blocks a certain number of photons from going through. Attenuators are rated in decibels according to the amount of photonic loss they elicit.

Once the wavelengths get to the termination device they need to be demultiplexed and sent to the appropriate receiver. The easiest way to do this is by splitting the fiber and shunting the same signals to all the receivers. Then each receiver would look only at photons of a particular wavelength and ignore all the others. This form of narrowband receiver is very expensive and inefficient to manufacture. Instead, most receivers are wideband, allowing wavelengths across most of the laser spectrum to be seen. However, only photons from a given wavelength should be allowed to get to the receiver.One method of stripping the desired wavelengths from the unwanted ones involves Bragg grating (see diagram). Bragg grating can be thought of as small lines etched across the core of the fiber. With the lines etched at specific distances from one another, wavelengths of light can be reflected back toward the input. A blazed grating is also etched into the core at an angle that reflects the light coming back on the grating out of the core and onto a receiver. By placing several combinations of Bragg and blazed gratings in a fiber, you can separate all the wavelengths out to waiting receivers.

Care and Feeding

Data would never get from Point A to Point B without the fiber itself. Fiber-optic cable comes in a variety of lengths with different connector types. Connectors are probably the most difficult part of the fiber-optic setup. That's because the names of the various connectors--FC, SC, LC, ST and MT--can be confusing, and remembering which one is which can be difficult. Most people remember connector types by giving them names that reflect how they work: FC (finger cramp, because fingers are easily cramped while trying to screw the connector tight); SC (stab click--the connector is stabbed into the socket until it clicks); LC (little click--a smaller version of SC); ST (stab twist--the connector is stabbed into the socket and then turned about a quarter turn to lock); and MT (mighty tiny--the smallest of the connectors). For more on connector types, see workshop, "10 GIG Can't Wait to Interoperate").

While today's fiber is more resilient than its predecessors, care should still be taken when handling it. Fiber should never be tightly bent or curved, and though you're used to walking on the copper pairs strewn around your computer room, fiber-optic cable should be treated like glass--don't step on it. Any change in the diameter of the core can cause great changes in dispersion, resulting in the fiber losing its transmission qualities.

The fiber-optic speeds are incredible. With today's fiber systems, you can transmit the entire contents of a CD-ROM in about one half of a second. And the technology is not standing still. The Optical Internetworking Forum is working to complete the interface standards for OC-768, which will increase bandwidth to 40 Gbps. That's the entire content of eight CD-ROMs, or 5 GB of data, transmitted every second.But even 40 Gbps is the tip of the iceberg. Alcatel continues to push the bandwidth on single fiber transmission by cramming 365 10-Gbps wavelengths over a single 6,800-km fiber. Designed for marine installations, this technology would transfer more than 3.5 Tbps, or enough bandwidth to carry 47 million simultaneous phone calls.

True optical switching is the next frontier, and companies like Opthos are already blazing a trail. Most switches today rely on OEO technology, with switching actually occurring within the electrical components. By switching at the optical layer, it occurs not only faster, but cooler and quieter as well.

Darrin Woods is a Network Computing contributing editor. He has worked as a WAN engineer for a telecom carrier. Send your comments on this article to him at [email protected].

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like


More Insights