Information Technology Reference
In-Depth Information
Cable discontinuities can be found by time-domain reflectometry techniques.
This applies to either copper or fiber. A substantial mismatch, in impedance or in
refraction, will clearly show up on most cable scanners with a graphic display. You
will be able to determine the distance from the scanner to the cable defect. Be sure
to subtract out the length of the test cord when you hunt for the problem.
Ordinarily, this test will direct your attention to a termination point, such as cross-
connect, patch, or outlet. Infrequently, you will find a problem within a cable run,
often where the cable has been severely bent and straightened out improperly.
Finally, lost performance can come from older fiber that just does not have suf-
ficient bandwidth to operate with gigabit at a given distance. For example, Gigabit
Ethernet can operate to 275 m at 850 nm on 62.5/125 mm multimode fiber that has
a bandwidth product of 200 MHz•km. However, the link is limited to an expected
range of only 220 m with 160 MHz•km fiber with the same core/cladding specs.
Workmanship.
Frankly, at lower speeds, you can be fairly certain that any link
that could pass the simple tests would also pass the full cable scan, as long as work-
manship practices were reasonable.
However, at gigabit, you must be perfect. A single pair (our favorite is the
brown/white) that is not twisted up to the point of termination can cause the link
to fail. Fortunately, the workmanship practices are not particularly different from
what they were for Fast Ethernet. You must actually follow them with gigabit. You
could get away with a few cable kinks, some sharp bends, and careless placement of
cable, but not with gigabit. It uses absolutely every ounce (or gram) of margin in a
Category 5 cable, and it cuts you no slack.
Be kind to your installers, though. Four or five years ago, cable installation
techniques were simpler. Many installations were accomplished with Category 5
components before all the testing criteria were determined.
As a result, many installations were tested only to Category 3, although higher-
performance components were used. In most of these installations, operation to Fast
Ethernet speeds is fairly well assured. But too much new information and practice
has appeared for anyone to have provided for gigabit performance at that time. This
is the reason the TSB-95 was created: to recognize that additional testing must be
done on legacy Cat 5 installations to ensure operation at gigabit.
At one time, speeds as great as 10 gigabits were considered almost science fiction.
Before the rapid acceleration of the Internet lit up network cores at amazing traffic
levels, there seemed to be little need for 10 GigE outside the lab. However, moving