March 26, 2001, 1:25 PM — Every child of the '70s remembers the first mainstream application of fiber optics: Color-changing lamps that were as cheap as they were mesmerizing. Nowadays fiber has to handle more serious but just as fascinating jobs.
Laparoscopic surgical instruments have transformed innumerable invasive procedures into outpatient surgery. Fiber cable bundles move telephone and TV signals through narrow ditches, helping rid streets of weather-sensitive poles and suspended wires. Major telephone companies' nationwide fiber networks supported the competition that sent long-distance rates into free fall. In corporations, fiber is the technology of choice for SANs (storage area networks) and for bridging distances between buildings and between floors.
In every sense, fiber optics is a remarkable technology. One day it will replace copper cable. Fiber advocates, including Senior Analyst P.J. Connolly, can't understand why all companies aren't already reworking their infrastructure to bring fiber to every desk. On the flip side, copper supporters, such as East Coast Technical Director Tom Yager, think those dollars and that effort would be better invested in solving other network problems.
Tom: I admit to being one of those kids who thought those chintzy fiber optic lamps were cool. And I'm no fiber Luddite now. For long hauls, for hazardous environments (such as where volatile gases might be ignited by sparks), and for massive storage networks, fiber is indispensable. In these cases and a few others, the electrical signals that travel over copper cable are a liability.
However, fiber's success in interbuilding hauls and SANs does not justify the pressure that's being applied to get companies to rip out their Category 5 network infrastructure and replace it with fiber.
Copper survived the transition from 100M bps to 1,000M bps, something a lot of fiber hounds said it couldn't do. Now I'm hearing that gigabit is the absolute ceiling for Category 5, and that any company that doesn't want its network to fall behind bandwidth demands had better start fishing fiber through the walls now.
I don't buy it, and I certainly don't advocate spending money on a fiber-to-the-desk refit when more and more of the people who occupy those desks are getting pink slips. It doesn't make economic sense.
P.J.: I prefer Lava Lamps myself, but I'll suggest other reasons for choosing fiber over copper.
For one, I'm surprised that I have to mention lightning to a Texan. Also, although it sounds like something from James Bond, copper networks are a source of radio frequency emissions. If someone is motivated enough, they can capture your traffic without ever touching your physical infrastructure.
Besides, the costs of implementing fiber are way overblown. Although you're probably going to pay a premium for stuff like patch cables and tools -- most desktops come with 10/100Base-T network interfaces built-in -- that's not where you're going to implement fiber anyway.
I agree that a general overhaul of your network isn't called for right now, unless you're still running unswitched 10Base-T on Category 3. But if you're going to be in your location for more than a couple of years, it's time to plan your move to fiber.
Most worthwhile investments don't make sense when you look at the short term. But if you can look past the recent stock market slide, you'll see that there's a lot to be gained from rethinking your entire cabling system.
Going comparison shopping
Tom: I saw one story that called the cost of fiber patch cables "comparable to that of Category 5." I did some shopping and found 15-foot fiber patch cables for US$34.95. If that's what the author of the story is paying for Category 5 patch cables, I'm resigning InfoWorld, and I'll send my kid to Princeton selling 50-cent cables to that guy for $35 a pop.
Today's gigabit copper equipment costs (excluding cable) are comparable to gigabit fiber. An Asante four-port gigabit copper switch retails for $1,399. That's only $100 less than the least expensive four-port gigabit fiber switch I found. If that pricing would hold, I'd join the analysts saying "if you're going to pay for gigabit, you might as well buy fiber."
But that pricing won't hold, not even for six months. When 100Base-T (100Mb copper Ethernet) was introduced, equipment was similarly expensive. Over the course of a year or so, prices fell below levels associated with 10Base-T gear.
I predict that gigabit copper prices will fall even more rapidly than 100Base-T did. The proliferation of 100Base-T was stymied while companies replaced the last of their coaxial cable with twisted pair. But most companies are already wired for gigabit copper.
