Published In July 2000
Everybody trying to sell you the latest technology claims his or her solution provides "futureproofing" that justifies its higher cost. Whether it's the fastest processor in a PC or the highest grade of copper cabling or fiber optics, nothing can futureproof a network for more than a few months or a year. I state this with confidence for several reasons. No one can predict the future when the present changes so fast. We can't even reliably predict the weather or certainly the stock market, so how can we predict changes in technology? Consider the pace of change in communications since the invention of the telephone. From 1876 to 1960, the telephone system was practically the same. Electronic switching became prevalent in the '60s and '70s, lasting 20 years before digital switching replaced it. Copper wire lasted 110 years before fiber optics replaced it in the long distance market. Now, new developments are being rolled out in fiber optics on a monthly or weekly basis and put into service immediately. Fiber to the home is looking feasible, too. The first 25 years of computers saw only marginal changes. It was not until transistors replaced the tube in the '60s that computer power grew and minicomputers appeared. Integrated circuits revolutionized computing in the '70s and led to the PC revolution. PCs have been doubling in power every 18 months for almost 20 years. PCs are obsolete by the time you get them out of the store, as new, more powerful machines are coming onto the sales floor from the warehouse! Computer networks are only 30 years old, but have grown from 1 Mbps in 1970 to 10 Mbps in the late '70s, 100 Mbps in the late '80s, 1 Gbps in the late '90s. That tenfold increase in speed per decade isn't holding, however, as 10 Gbps local area networks (LANs) are only going to be between three and four years behind the 1 Gbps introduction. Technology doesn't always obsolete what exists "Techies" get used to change and assume each new generation of technology will obsolete the last. Sometimes, these technological changes allow us to preserve the past. Consider, for example, the automobile with its internal combustion engine. When air quality concerns led to automotive emission controls, many predicted that the internal combustion engine would die and be replaced by electric motors, turbines, or the Wankel. (Remember the Wankel rotary engine?) For a few years it looked like the internal combustion engine was a goner. Emission controls ruined its driveability and fuel economy. But electronic engine controls and fuel injection, coupled with high-tech analyses of the combustion processes, has given new life to the internal combustion engine, even for those who'd rather have a V8. Likewise in the voice/data network business, technology has given new life to old equipment. Digital subscriber line (DSL) electronics have enabled old telephone wires to be used to send high-speed digital signals. Sophisticated digital signal processing makes current Cat 5 copper cabling capable of speeds up to 1 Gbps. And optical advances are making multimode fiber optic cable capable of 10 Gbps. Lets look at two situations in the voice/data cabling business today that illustrate my points-the last two points mentioned above. Cats 6 and 7 cabling systems considered "pre-standard" I recently attended a seminar given on a manufacturer's latest generation cabling scheme. They had gone way beyond Cat 5 and enhanced Cat 5 (Cat 5e) to offer "Cat 6" and "Cat 7" cabling systems. Cat 5e has just been approved by the EIA/TIA TR42 committee, which writes the structured cabling standards. Cat 6 is in the proposal stage, so manufacturers refer to it as a "pre-standard" cabling system. Practically every cabling system manufacturer sells cabling systems they call either "Cat 6" or "Cat 7." The tester manufacturers tell everyone that these cabling systems are incompatible, each requiring custom test interfaces for the tester to compensate for their differences. Many manufacturers sell their system on the basis of "futureproofing" your cabling system and give you a long warranty on its performance (It is usually full of fine print!). But if these cabling systems are incompatible, it's highly unlikely that future networks will work on them, or on more than one, which means you are gambling that you pick the right one! The manufacturer giving the seminar I attended had the only legitimate reason for installing its higher-performance cabling solution; it offers more "headroom" for today's networks. So maybe their expensive cabling can compensate for less-than-ideal installation practices, but is it worth the premium in cost? High-performance multimode fiber The fiber manufacturers are selling two new solutions for higher-bandwidth fiber optic cabling, which provide greater distances for Gigabit Ethernet and the possibility of running 10 Gigabit Ethernet in the future. The two solutions are 62.5/125 fiber with higher bandwidth for 850 nm lasers (called VCSELs) and 50/125 fiber. Both have a small premium in cost over today's standard 62.5/125 fiber, but the 50/125 fiber is also incompatible with practically all fiber installed in the United States in the last 15 years. Both options promote their advantages for Gigabit Ethernet, a legitimate concern for LAN backbones but of doubtful usefulness for fiber to the desk. And 10 Gigabit Ethernet is purely in the research and development (R&D) stage, years away from real systems. Furthermore, the VCSEL technology that gave us low-cost lasers at 850 nm is now yielding 1,300 nm lasers, compatible with single-mode fiber. Single-mode fiber, which is standard for telephone companies and CATV companies, offers practically unlimited bandwidth and is less expensive. The combination of 1,300 VCSELs and single-mode fiber may obsolete multimode fiber soon. Hire a fortuneteller? How do you make decisions in such turbulent times for technology? Limit your view of the future to three to five years and the price premium you are willing to pay for "headroom" to 10 to 15 percent. All "pre-standard" solutions are a gamble, and you are betting on one vendor's vision of the future. The only thing we know about the future is that we are terribly inaccurate at predicting it! HAYES, a frequent contributor to Electrical Contractor, is president of Fotec in Medford, Mass. He has provided fiber optics training and written widely on fiber optics, including The Fiber Optics Technician's Manual. He can be reached at email@example.com.