Free-Space-Optics-Wireless-Without-the-Radio

I mentioned earlier that EHF airlinks can rather easily be paralleled with free-space optical
(FSO) transmission systems. Both FSO and EHF are vulnerable to adverse weather conditions—
FSO to heavy fog and EHF to rain and snow—but, on the other hand, EHF is not much
troubled by fog, and FSO is not obstructed by raindrops. Thus, used together, the two systems
can achieve an availability figure that is at least an order of magnitude better than either alone
over a given distance. And because the normal operating distances of the two technologies are
similar and because both transmit tightly focused line-of-sight beams, the two work well in
tandem.
At the same time, FSO may be considered to be a competitive technology with respect to
802.16 millimeter microwave. It addresses essentially the same market segments and shares a
similar freedom from wireline infrastructure. It is also akin to millimeter microwave in that it is
a young technology, appearing in commercial form in the 1990s and failing to establish a large
market presence thus far.
In at least two other respects, however, the two access technologies are quite divergent.
FSO has the potential to offer significantly higher throughputs than EHF will ever achieve.
Many of the same techniques used in fiber-optic networks, such as dense wave division multiplexing
(DWDM), subcarrier multiplexing, 40GHz modulators, superfine optical filters, and
optical Code-Division Multiple Access (CDMA), can also be used in FSO systems, and terabit
speeds are theoretically possible over short distances. Indeed, some years ago, Lucent
announced it had achieved 80 gigabits per second (Gbps) transmissions in the laboratory using
DWDM alone. Surely that figure can be bettered in time. However, microwave throughputs are
likely to achieve only incremental gains, and even these will require fundamental advances in
high-speed devices and modulator circuits rather than building on technology that already
exists, which is the case with FSO.
We must keep in mind, however, that terabit speeds or even large fractions thereof have
not been demanded of any metro backbone except a lateral access connection extending from a metro hub. The great error of the telecom bubble was to assume that demand for bandwidth
was insatiable and would drive the rapid adoption of faster and faster access technologies
regardless of cost or other limitations. In fact, carriers have not found a great deal of demand
for optical wavelength services in the range of 1Gbps to 10Gbps. It remains to be demonstrated
whether demand would be more intense for a wireless service of comparable speed.
The other point of divergence between FSO and millimeter microwave has to do with
distance. While some manufacturers of FSO equipment have claimed transmission distances
of up to several miles, no commercial deployment yet shows that. Sending an infrared signal
over great distances is not infeasible, but the power levels necessary to permit long-distance
links render the transmitter unsafe to birds and to the eyes of any animal including Homo
sapiens unlucky enough to blunder into the path of the beam. Higher power levels are also had
at the price of throughput since laser modulators suffer the same trade-offs in terms of
frequency and power level afflicting microwave output transistors.
To date, FSO systems have been costly, high maintenance, and critically short in useful
range. Costs are coming down, and, at the same time, auto-alignment systems are reducing
maintenance requirements substantially. Distance limitations will not be so easily solved,
though, and these limitations will confine FSO to niche applications for the foreseeable future.