On the Bandwidth Wagon

By Jim Hayes | Sep 15, 2005
generic image




You’re reading an outdated article. Please go to the recent issues to find up-to-date content.

It is common knowledge that fiber optics has much more information-carrying capacity than copper wiring, mostly due to the fact that fiber optic suppliers have been touting the bandwidth advantages of fiber for so long. The usual pitch is to install fiber optics and forget about the frequent upgrades you face with copper cabling. Few suppliers, however, have expected the rapid developments in network speeds that we have experienced in the last few years.

Granted that one generation of fiber optics—the FDDI-grade 62.5/125 micron multimode fiber used for virtually all installations for the last 20 years—has outlived coax, shielded twisted pair, and four generations of unshielded twisted pair cables. But network speeds have risen from 100 megabits per second five years ago to 1 gigabit per second today. Ten gigabits per second links are now being deployed in backbones and storage area networks. The high end of today’s networks is beyond the bandwidth capacity of some legacy multimode fibers.

Today’s users are probably best advised to install 50/125 laser-optimized fibers for their capability to handle gigabit and 10-gigabit networks in the future. But what about all that fiber installed in buildings during the last 15 or 20 years? Will it support these high-speed networks? Until recently, there was no answer. With no way to test these older fibers, we had to rely on simulations and historical data.

Once telco applications of single-mode fiber took off in 1983, the fiber companies focused all their R&D on the single-mode version. Multimode fiber was still manufactured, but since it represented only a few percent of the whole fiber market, development effort was minimal, mostly trying to make it cheaper.

The 62.5/125 multimode fiber that became the industry standard was adequate for LANs and short links based on LED transmitters for almost 20 years. But once gigabit LANs were developed, LEDs had to be abandoned as they were too slow. Low-cost 850 nm lasers called vertical-cavity surface-emitting lasers (VCSELs)—which could be manufactured cheaply like LEDs—offered the performance needed by gigabit links.

FDDI-grade 62.5/125 micron fiber had relatively low bandwidth at 850 nm and its design and manufacture was sometimes not optimal, especially for laser sources. The graded-index profile of the core was sometimes not uniform, creating problems with modal bandwidth for lasers. The bandwidth performance of 62.5/125 fiber became the limiting factor in the distance gigabit Ethernet could be supported in networks.

Fiber manufacturers reached back into the parts bin and pulled out the “latest” 1984 version of 50/125 fiber, which had high bandwidth at both 850 and 1,300 nm. It turned out to offer better performance (i.e., longer links) than 62.5/125 fiber at a lower cost. And with some research and development, it was further optimized for 850 nm lasers (without any less performance at 1,300 nm). That is the fiber that users should probably be installing in new installations, but what about the old fiber already in the building? Will it support gigabit and faster networks?

That question needs an answer, and when scientists and engineers attack such a need, a solution is not far behind. A valid answer requires two steps: testing some of this old fiber, then trying to run high-speed networks on it. I’m pleased to say I have gotten involved in just such a project and useful data should be forthcoming.

Currently, I can only offer some preliminary data from a field test I attended recently. Using a prototype instrument at a government facility, we tested two old fibers representative of the 62.5/125 FDDI-grade fibers and 50/125 fibers installed all over this facility. The FDDI-grade fiber showed both limited bandwidth and pulse-distortion characteristics of fibers with some manufacturing deficiencies, indicating potential problems with gigabit networks. The older 50/125 fiber, however, looked quite good, fully capable of handling gigabit networks or more.

With the millions of meters of multimode fiber installed in buildings over the last 20 years, being able to test it for today’s high-speed networks is a big advantage for both the end-user, who needs to know if the present cabling is still useful, and the contractor, who can offer testing services or installation services if new fiber is needed.

Now may be the time to discuss this with your fiber customers. Are they concerned about using current fiber optic cable plants with future high-speed networks? Do they want to upgrade their current fiber-to-laser-optimized fibers or do they want to test their cable plant for its ability to be upgraded? The latter could be less expensive, but both require the services of experienced fiber optic contractors. EC

HAYES is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at


About The Author

HAYES is a VDV writer and educator and the president of the Fiber Optic Association. Find him at

featured Video


Vive Pico Wireless Remote

The Pico wireless remote is easy to install, it can be wall-mounted or mounted to any surface, and includes a ten-year battery life. See how this wireless wall control makes it simple to add lighting control wherever you need it.


Related Articles