The impending fiber-bandwidth shortage

How often have we heard that fiber has "unlimited bandwidth," or that if you install fiber today you will never have to recable because it can handle any application? Such misconceptions are shortsighted attempts to convince the industry that there is a single solution to bandwidth needs. In reality, fiber suffers from the same bandwidth constraints as unshielded twisted-pair (utp) cable--as application speeds increase, the data throughput decreases.

The Imending fiber-bandwidth shortage

Scott Stevens

How often have we heard that fiber has "unlimited bandwidth," or that if you install fiber today you will never have to recable because it can handle any application? Such misconceptions are shortsighted attempts to convince the industry that there is a single solution to bandwidth needs. In reality, fiber suffers from the same bandwidth constraints as unshielded twisted-pair (utp) cable--as application speeds increase, the data throughput decreases.

When utp cable faced a similar problem in the late 1980s, many in the fiber industry insisted that 20 megabits per second was the top end speed for utp. If you ever wanted to reach speeds of 100 Mbits/sec, you would have to use fiber, they said. But with utp, manufacturers design in the bandwidth performance. By engineering a better cable, utp has easily reached 100 and 155 Mbits/sec, and will soon support gigabit applications (albeit over only 100 meters). The process is ongoing. utp manufacturers continue to improve the performance of their cables and the ability of those cables to handle higher data throughput rates.

Contrast this with fiber optics, where the bandwidth is locked in at the time the fiber is drawn. To improve the transmission capability of fiber, the glass manufacturer must improve the product. Yet while utp has evolved, fiber performance has stayed the same. Today`s standard-grade 62.5/125-micron fiber has the same bandwidth it had 10 years ago--160 megahertz/kilometer at 850 nanometers and 500 MHz/km at 1300 nm.

While some may contend that maintaining a single standard of performance is good, the problem is that the applications have begun to outpace fiber`s ability to handle the data. Although the eia/tia-568a commercial building telecommunications wiring standard and bicsi`s Telecommunications Distribution Methods Manual have assumed a 2-km design distance for standard-grade multimode fiber, the soon-to-be-released Gigabit Ethernet standard ensures operation only to 260 meters at 850 nm and 440 meters at 1300 nm. Campuses with multimode backbones greater than 260 or 440 meters may not be able to support the new gigaspeed applications.

And this is only the start. It would again be shortsighted to assume that development of higher-speed applications will stop at 1 Gbit. Speeds between 2.4 and 10 Gbits/sec will be developed soon, and fiber`s constraints will become even more noticeable. Even interbuilding backbone cabling will be affected, and offices with installed fiber plant will have to be recabled to support emerging technologies. Already eia/tia telecommunications systems bulletin tsb-72, "Centralized Optical Fiber Cabling Guidelines," is outdated. It calls for maximum distances within a building of 300 meters, which is 40 meters beyond the limits of Gigabit Ethernet at 850 nm.

What`s the solution? Singlemode fiber is an obvious choice for the long term, though it is an expensive option. Furthermore, it cannot support the inexpensive short-wavelength technologies such as 10-Mbit/sec Ethernet, so it does not provide an optimum migration path. A second alternative is to improve the performance of multimode fiber. Some manufacturers can provide bandwidth performance of up to 1 gigahertz/km at 1300 nm as a standard offering--a 100% increase over generic fiber performance. The improved bandwidth means better performance at higher speeds over longer distance.

This brings us back to the utp analogy. No one would run high-speed data over Category 3 cable today because more-advanced copper cables are available. One could use Category 3, but the maximum transmission distance would be very short. The same applies to fiber. A 500-MHz/km fiber could be used to run gigabit speeds, but only with significant distance limitations. A better approach is to find manufacturers willing to produce cable with better performance. Users who want to ensure that their investments continue to pay dividends would be well served to install the highest-bandwidth cable available.

Scott Stevens is technical services man-ager with CommScope (Claremont, NC).

More in Home