Addressing the Bandwidth Problem in the Lower Microwave Regions

It should be noted that some radios do permit simultaneous operation in two different bands,
significantly increasing the amount of bandwidth available to the network operator and ultimately
to the subscriber. Most commonly, dual-band radios will use one band for the uplink
and another for the downlink. One may, for instance, use the 2.4GHz unlicensed band for an
uplink and the MMDS licensed bands (2.5GHz to 2.7GHz) for a downlink. Downlinks, it should
be understood, normally utilize far more bandwidth than uplinks; in other words, the network
is asymmetrical—the theory being that most residential and small business users tend to
download far larger files than they transmit, and that the Web browsing experience will be
greatly enhanced by a faster downlink as well.
It is possible to design radios that can span more than two bands; one could, for example,
build a radio that could operate in the lower ISM band (902MHz to 928MHz), the unlicensed
band centered at 2.4GHz, and all three of the three unlicensed bands between 5GHz and 6GHz.
A company called Wireless Inc. put out just such a product a few years ago but was unable to
find a market for it and ceased production. The basic concept remains valid, however, and
I predict that as demands for bandwidth grow among 802.16a network operators, it will be
revived.
In the midterm, radios will begin to appear that are not bound to fixed bands at all and will
enable the network to use almost any frequency desired; that is, the radio would not have to
be factory tuned to just a few bands but would have the flexibility to select any band and could automatically accommodate itself to changes in bandwidth allocations on the part of
governing bodies. Termed frequency-agile radios or software-defined radios, such products will
in most cases be able to select a modulation system as well.
Software-defined radios exist today, and at least a score of companies are active in this
area, but no one is making a WiMAX-certified product conforming to the 802.16a standard, and
in fact most sales to date have been to military organizations. The majority of software-defined
radios made thus far have utilized costly field programmable gate arrays (FPGAs) and have not
met the price requirements of network operators. Incidentally, several have combined frequency
agility with an adaptive array antenna using the extreme processing power provided by
the FPGAs to perform multiple functions.
Frequency agility is unquestionably desirable in governmental applications. It provides a
radio operator with a means of concealing a transmission and evading jamming, and it allows
safety personnel to contact all relevant agencies over whatever frequencies have been allotted
to them. But in a commercial setting where relatively few bands can be legally occupied by any
individual network operator, the need for such flexibility is unclear at present. Many advocates
of software-defined radios look forward to a day when radio spectrum is brokered and made
available to consortia of network operators as needed, with payments automatically being
made to license holders or with bandwidth being swapped among participating networks
within a peering relationship. Such a grand schema would certainly result in much more efficient
use of spectrum than is presently the case, but it is difficult to envision how multiple
networks with countless users could all coordinate their transmissions so as to avoid interference.
Everyone would have to be equipped with an intelligent radio, and each radio would
have to connect to an overarching computing grid devoted to the management of all available
spectrum.
I will not speculate further as to how or even if such a virtual organism could come into
being, and I do not see even the beginnings of such a regime occurring within this decade.
Nonetheless, some degree of frequency agility will manifest itself in the lower microwave
regions within five years, and most likely it will involve subscriber terminals utilizing 802.11,
802.16, and 3G mobile networks as the situation warrants and in accordance with roaming
agreements among a relatively few incumbent carriers.
Software-defined and frequency-agile radios will undoubtedly influence both spectrum
use and the posture of regulatory bodies over time, and they will greatly alter the situation of
network operators by making significantly more spectrum available to them, perhaps in
amounts approaching 1 gigahertz. Eventually the increase in available bandwidth will expand
the uses to which the network may be put and will open the possibility of the network operator
offering a wide range of high-fidelity multimedia content as well as simple high-speed access
and LAN extensions. At the same time, I caution broadband wireless operators today not to
predicate their business plans upon the imminent arrival of such technology in the marketplace.
That may be years away, and in the meantime operators will have to adapt their business
to relatively scanty bandwidth.
I will now add a final word on current throughput constraints in the lower microwave
regions and how they may be overcome in the future by technologies other than softwaredefined
radios and the related adaptive array antennas. Over the course of the next several
years, radios will steadily increase their ability to resolve signals in the face of greater and
greater amounts of noise and interference and in their ability to reconstruct information that
has dropped below the noise floor. These abilities will allow operators to pack more and more
information into a signal by increasing the number of discrete phase and amplitude states used to convey information and, in the case where subcarriers are employed, decreasing
the spacing between them. All of this will lead to higher throughput and, combined with the
aforementioned technologies of adaptive antennas and software-defined radio, may eventually
allow the achievement of fiberlike speeds within the lower microwave region. Fiber
itself will not stand still, however, though it remains to be seen what applications will emerge
that will require throughputs in the tens of gigabits per individual users.
In any case, the midterm future of wireless broadband promises to be an era of very abundant
throughputs and greatly expanded service opportunities for the operators. Five-year
market projections should consider the likelihood of fundamental improvements in the core
technology and should not rest upon the assumption that the business will remain much as it
is today. The operator should take it almost as a given that generations of equipment will be
short and that capital improvement will have to be undertaken on an ongoing basis in order
to remain competitive. For the foreseeable future, the low microwave wireless broadband
network is not going to be a set-it-up-and-forget-it cash cow.
In the current regime characterized by throughput rates that are generally competitive
with those of cable and DSL, and by network capacities that are generally much lower, network
operators must proceed cautiously, however, recognizing that they lack the resources to be all
things to all customers. Their aim must be to utilize the capacity of the network as efficiently as
possible and target those customers who will bring them the best return on the infrastructure
investment.
This aim cannot be achieved in most markets simply by presenting subscribers with a
“pipe,” or raw capacity. Raw capacity to support broadband access is becoming increasingly
available and increasingly commoditized, and the mere fact that a wireless carrier is offering
capacity is scant inducement for the customer to subscribe. Wireless broadband represents an
opportunity for competitive access provider, perhaps the only opportunity to enter many
markets, but is not inherently highly attractive to end users simply by virtue of being wireless.
If the wireless operator is competing with well-entrenched wireline incumbents who are
prepared to wage price wars, and if that same wireless operator is offering nothing more than
basic services, his prospects of succeeding are poor because the service offering is just another
broadband choice and one that is in certain respects technically disadvantaged vis-à-vis
the others.
Chapter 2 mentioned the major value-added services that can be supported over wireless
networks. The following sections attempt to be more specific with respect to the lower microwave
region.