Information Technology Reference
In-Depth Information
In many ways, the apparent speed and extreme cable lengths of these primitive
computer connections created the set of expectations for modern cabling. To this
day, centralized fiber-optic cabling (at speeds below gigabit) is the only networking
method that approaches old coax-terminal connection distances. Fortunately, the
practice of placing remote terminal servers (controllers) in each workspace coinci-
dentally limited the run length from controller to terminal. Even in a large building,
a controller was rarely over 200 cable-feet from its terminals. This has made the
subsequent 90-m TIA limit much more palatable to the mainframe manager, when
forced into the role of overseeing a distributed computer network.
In addition to the coax terminals, some computer installations used a new type
of computer-terminal serial data interconnection often referred to by the applicable
standard, RS-232. These connections required multiconductor cable, rather than
coax, and more severely limited the distance to the terminals. Eventually, we found
that you could easily exceed the recommended 50-ft operating limit of RS-232 by
utilizing “special” cable with twisted-pair stranded wires. In fact, the similarities to
telephone wire were inescapable. Network training at this point included knowledge
of which pins of the 25-pin interface to use, how to connect them in the proper pat-
tern (including the crossover or “null modem”), and how to arrange the pairing of
the unbalanced interface to maximize distance. Using the right techniques, you
could easily get 200 to 300 ft (about 60 to 90 m) reliably.
At about the same time, distributed computer networks arose, with the advent
of coaxial Ethernet. Actually, early Ethernet was a real bother to install. The coax-
ial bus topology required a thick and bulky coaxial cable to be run from each work-
station to another, and then required an equally thick AUI cable from the coax tap
to the workstation. Not only was it difficult to deal with, but an inexperienced
installer could create not-so-subtle network problems, from flaky connections to
total network outages.
Ethernet made the shift to thin coax in the mid-1980s at about the same time
we were beginning to do RS-232 over twisted pair. Naturally, transmission engi-
neers soon learned how to convert Ethernet's unbalanced thin coax signal to a bal-
anced twisted pair, and an industry was born. Thin Ethernet was also unreliable,
whether it was true coax or was converted to twisted pair by balun transformers.
However, the twisted-pair concept was a natural path to structured wire concentra-
tion in telecommunications rooms. Several manufacturers began to insert active
hubs into these networks, located in the telecommunications room.
Finally, the first modern hub-and-spoke, or star, networks were a reality, and
network wiring technology began to get serious. Standards were developed for
structured cabling, originally by the Electronic Industries Association (EIA), and
later by a subgroup, formed from an amalgam of industry groups and named the
