VanLUG Email Archive
Hmmm. Perhaps I *should* give that presentation on the effects of
propagation impairments on the bit error performance of fixed wireless
links at a future VanLUG meeting. And clear up a few misconceptions in the
process.
The fact is, a wireless channel is every bit as good (and fast) as an
equivalent wired channel if the bit error rate (BER) is kept low. We keep
the BER low by keeping (1) the Path Loss low, (2) the Ricean K-factor high,
(3) the Doppler rate (low), and (4) the delay spread low, at least relative
to the symbol duration. (It also helps to minimize turn-on delays in the
transmitter and to select appropriate coding, modulation, and, if
indicated, diversity schemes, but I digress.)
The absolute worst propagation environment is between a unit mounted in a
fast moving vehicle and a base station located somewhere in the urban
jungle. Here, Path Loss is frequently high, the Ricean K-factor is
virtually zero, the Doppler rate is high, and the delay spread is high
(because we're using omnidirectional antennas). It simply doesn't get much
worse than that! In such cases, BER is high, retransmissions are frequent,
and throughput is poor.
The absolute best propagation environment is between fixed units with a
clear line-of-sight between them. A cloth roof, such as the roof of Canada
Place, is almost completely transparent to RF and introduces little, if
any, additional path loss. In such cases, BER is low, retransmissions are
rare, and throughput is high.
Hint: Which case corresponds to a link between Canada Place and Mt.
Seymour?
In practice, the performance of a 56 kbps radio link over such a path would
be comparable, and perhaps even faster than a connection established using
an analog telephone modem.
However, once other issues are considered, it might be best to go along
with the show management and simply rent a phone line for three days.
-- Dave Michelson