Optical fiber, a glass wire thinner than a human hair designed to transmit light, is one of the true wonders of our age. One recent claim holds that a single fiber strand will eventually be able to transmit 2 trillion bits per second; the total transmission of the entire telecommunications industry today...
Services   Books   About   New   Contact   Home  

Every Manager's Guide to Information Technology:

Extract (6): Fiber Optics

Optical fiber, a glass wire thinner than a human hair designed to transmit light, is one of the true wonders of our age. One recent claim holds that a single fiber strand will eventually be able to transmit 2 trillion bits per second; the total transmission of the entire telecommunications industry today, including AT&T, MCI, US Sprint, and the Bell Operating Companies, is only 1 trillion bits per second. At present, typical fiber speeds are in excess of a billion bits per second. The SONET standard operates at up to 2.5 billion bits per second. An early version of SONET was introduced in Chicago in early 1991. MCI announced at the end of 1993 plans to build a $20 billion network to bypass the local phone companies, which charge MCI more than $5billion a year in "access charges." This new network will initially serve 20 major cities, and SONET will be the key. MCI could never afford to replicate the Baby Bells' massive existing infrastructure of copper cables, but it won't have to; it can install fiber in the old telegraph lines, some dating back a century, that MCI obtained in its acquisition of Western Union in 1990. MCI can then transmit voice communications at gigabit speeds.

It will take time for telephone companies in the United States and abroad to replace all existing copper cable with fiber, but the process has begun. Transatlantic telephone capacity doubled with the installation of the 8-fiber strand TAT-8 system. Fiber optic-based local area networks are becoming increasingly common, with the FDDI standard (Fiber Distributed Data Interface) providing speeds of 100 million bits per second.

Fiber transmission speeds do not translate directly into transaction processing speeds any more than aircraft speeds translate into travel times. Airport congestion and delays, check-in, and road traffic all extend total journey time. Their equivalents in communications are the switches that route traffic onto the backbone network, which are not yet fast enough to take full advantage of fiber speeds.

Apart from speed and cost, fiber optics offers several advantages over cable among them, reliability and security. Fiber links, unlike cables that transmit electrical signals, cannot be tapped. The vulnerability of fiber is that if the hair-thin strand is cut, transmission ceases. United Airlines, realizing that the fiber strands that carry 100,000 simultaneous transmissions are usually buried only a few feet underground, dispatched a team to post the locations of some of its own fiber cable. Unfortunately, one of the signs was driven through the cable it was to protect, knocking out the airline's reservation system and halting marketing and sales for several hours. A British cable manufacturer had a similar experience; it planted a Christmas tree through the main cable of its head office building. Every large company is vulnerable to such chance interruptions of service. The more of its fiber's capacity a company uses, the more it will need to provide redundancy and backup linkages and establish recovery procedures. The prospect of simultaneously interrupting 100,000 telephone conversations, and the cash flow and customer service of any number of businesses, is daunting, indeed.

The billion-bits-per-second speeds of fiber may seem like a solution looking for a problem. Business communications today largely operate at speeds from 56 to 64 thousand bits per second. What will firms do with the extra capacity fiber affords? First, they will share it, by "multiplexing" many low-speed transmissions onto the fiber. They will also use it to move large data bases between locations electronically. Most important, they will use it for novel video, image, and engineering applications.

An example of what fiber might make practical and cost-effective is interactive access from a computer-aided design workstation to a full-color, full-motion, detailed design simulation of car performance running on a supercomputer. To provide the picture quality of a good photograph with the quality of movement on television or film requires a resolution of a megabit of pixels (the tiny dots that make up an image on a monitor display) per frame and a transmission speed of 720 megabits per second. To provide full-color, multiply these requirements by 24 bits per pixel. To ensure full motion quality, the images must be sent at 30 frames per second (the same speed used in projecting films). Clever tricks of data compression can reduce this to 15 megabits per second. Such an application is utterly impractical over copper cable across long distances.

In late 1994, the relative costs of copper cable versus fiber were (for 36-user workstation "drops," or connections):

cable: $3,000 for materials plus $2,400 for installation labor

fiber: $5,500 for materials plus $2,400 for installation labor

This works out to $150 per drop for cable versus $220 for fiber, a 50 percent differential.

Copper, however, is by no means the telecommunications equivalent of the buggy whip. Improvements in transmission techniques can now match FDDI's fiber optics standard that operates at 100 megabits per second, versus the 1980s' high-end local area network capability of 16 Mbps. Copper Distributed Data Interface (CDDI) can match that 100 Mbps speed. Bell Atlantic's local phone system delivers pay-per-view movies at 1.5 Mbps, using the same twisted-pair cable you often untangle when talking on the phone.

But fiber is the future. Experts estimate that installing a complete national fiber optics system in every home will cost about $200 billion and take at least until 2015 to complete. New Jersey Bell has committed to make its state the first to have fiber in every customer's location by 2010. Southern New England Telephone expects to connect 40 percent of Connecticut by 1997.

All this lengthy fragmented, but substantial, innovation is part of the so-called Information Superhighway, less a network than a network of networks of networks, provided by government, long-distance, and local communications providers and universities, cable companies, and business consortia. While no one knows how it will evolve, we do know it will be digital and fiber based.

Order from: Amazon Order from: Barnes & Noble