Deeper into Point-to-Multipoint

Point-to-multipoint architectures, as you have seen, are the norm in pervasive broadband
metro deployments and always have been. The success of such deployments depends, on the
one hand, upon reaching as many potential subscribers within the area swept by a single base
station and, on the other hand, on utilizing the available spectrum efficiently and effectively.
To illuminate how both objectives may be achieved, you must first examine the concept
of wireless coverage areas, commonly referred to as cells, and how they figure in a point-tomultipoint
architecture. Cells are the basic building blocks of wireless networks, and the mapping
process of planned cell sites within a given territory to be covered constitutes the most
basic strategic planning function of the network engineer.
Radio Cells and What They Portend
The concept of cells appears to have originated at Bell Labs, the research arm of the Bell Telephone
system, back in 1948. At the time no effective means existed for automating the tuning
function of individual radios to enable the concept to be realized, but later—30 years later, in
fact—the Swedish telecommunications giant Ericsson would successfully demonstrate the
feasibility of the concept in the first commercial cellular telephone system set up in the city
of Stockholm.
Today cellular architectures are ubiquitous in wireless communications, not just in
cellular telephone systems, but in wireless local area networks (WLANs), personal area communications
networks such as Bluetooth, and in fixed-point broadband wireless networks of
the sort constituting the subject of this book.
The Function of Cells in the Point-to-Multipoint Network
To grasp fully the cellular concept, one must first understand how traditional radio networks
operate, for the cell is a fairly radical departure from older practices and, at the same time, can
really only be understood in their context.
For as long as radios have been used, and they have been used for more than 100 years
now, the typical approaches in two-way network communications have been either peer-topeer
linkages where one radio transmits to another, as in a citizens band radio network, or
arrangements where all of the individual user terminals go back to one central base station.
The latter approach typifies the wireless telephone systems predating the cellular networks,
the two-way dispatch radios used in commercial fleets, and police and public safety radios. It also typified the first attempts to deploy commercial broadband data networks in the
United States.
The limitations of simple peer-to-peer, citizens band radio being a prime example, as
opposed to the similar but much more sophisticated mesh, are fairly obvious. Each individual
radio can reach adjacent radios only if it is not to broadcast interference far and wide, so such
an architecture cannot form the basis of a high-speed access network.
The limitations of the second approach are not so obvious, but they are nonetheless real.
Simply stated, a single base station has only a certain amount of spectrum to utilize to reach all
the subscribers in a given geographical area. Thus, in early mobile telephone systems where at
most a few dozen channels were available, a few hundred subscribers were all the network
could support using a widely accepted ratio of ten subscribers to every one of the available
channels. (Such a ratio is based on the concept of statistical multiplexing, where statistically
there may be only a 10 percent chance of any one subscriber transmitting at any given instant.)
Obviously, a mass market service such as cellular telephony must support not hundreds,
but thousands or even tens of thousands of subscribers within a metropolitan area, so it
demands something more, either a vast amount of spectrum or some way of reusing a more
limited number of channels so that more than one user can occupy a given channel simultaneously.
Given the total demand for spectrum on the part of a huge array of powerful interests,
the vast amount option is not really available; at least it was not until the emergence of ultrawideband
radio recently. The only choice the operator had was to find some means of reusing
spectrum, and the cellular approach, described next, spoke to that need.
So here, in brief, is how cells figure in a radio network and permit extensive frequency
reuse: Radio cells themselves are relatively small areas surrounding base stations. Both the
base station and the subscriber terminals within the cells transmit at very low power so that the
signal quickly fades out to the point where it will not interfere with someone else on the same
channel a fairly short distance away. Thus, each cell becomes in effect a subnetwork.
In principle this is simple enough, but, for the scheme to succeed, one has to work with
quite high frequencies that will fade quickly over distance, and during the first two decades following
Bell Labs’ enunciation of the principal of cellular reuse, the only devices that could
transmit at such frequencies were exotic vacuum tubes such as klystrons and traveling wave
tubes, neither of which was suitable for inclusion in a subscriber terminal. Moreover, if one
wanted to use cells in a mobile network, further problems presented themselves because a terminal
in motion would always be passing out of range of particular base station and had to
have some means of instantly acquiring an unoccupied channel in an adjacent cell if a transmission
was to be maintained. Such “handoffs,” as they have come to be known in the cellular
industry, require the use of powerful computers and specialized software at the base station
and some computing ability in the terminal itself. The microprocessors that could support
such functionality in a handset were not available much before the opening of the first network
in 1978.
Since cellular telephones established themselves in the early 1980s, the principle of cellular
reuse has also been applied to WLANs and internal wireless phone systems called wireless
PBXs and of course to fixed broadband wireless metro networks. In the case of the latter, handoffs
are not ordinarily required, so the design of the network is somewhat simplified over that
of a mobile network, but the basic design principles are much the same, as are the benefits,
chief among them reduction of transmit power requirements for terminals and vastly
increased spectral efficiency over the entire network because of aggressive frequency reuse. I point out, however, that the cells deployed within broadband wireless networks are usually
larger than those in cellular telephone networks.
A Note on Channel Assignments from One Cell to Another Concerning this matter of frequency reuse,
it should be understood that channels, according to standard engineering practice, cannot be
reused from cell to cell because a signal fades only gradually over distance and will not be sufficiently
attenuated so as not to cause grave interference if another user tries to occupy the
same channel in an adjacent cell. Therefore, a channel would normally be reused only one cell
diameter away at best. Figure 5-1 shows frequency reuse patterns in a cellular network; for
simplicity’s sake, it displays only four channels. The exception would occur when the cell was
divided into sectors, which are discussed next. Today this standard engineering practice has been subject to some modification because
of the appearance of advanced modulation techniques such as direct sequence, Code-Division
Multiple Access (CDMA), and orthogonal frequency division modulation (OFDM); the introduction
of polarization diversity; and the emergence of smart antenna technology. All these
techniques, described in detail later in this chapter, increase the immunity of a transmission
from interference and allow reuse of a channel at reduced distances—in other words, less than
one cell diameter away. Reuse, at least within certain sectors in adjacent cells, does then in fact
become possible, though certain minimum distances still have to be maintained. Coincidentally,
simple rules of thumb for channel spacing become increasingly difficult to formulate.
As it happens, numerous schemes and formulae have been developed for cellular telephone
networks for optimizing frequency reuse patterns, but all of these, understandably, are
optimized for mobility and must take into account the fact that a subscriber terminal must be
able to “see” two base stations simultaneously at the boundary of a cell in order to initiate a
handoff. As a result, such techniques cannot be imported into the broadband fixed wireless
arena without extensive modification. Considerably less attention has been given to maximizing
frequency reuse in fixed broadband networks through improved calculations for base
station placement or through new protocols for dynamic channel assignment.
Part of the problem has to do with the ad hoc nature of most fixed wireless broadband
networks to date. Typically, wireless broadband networks have been the product of startup
companies with limited resources that have tended to add capacity as needed with no overall
plan of what a fully loaded network would look like and with no staff with either the training,
inclination, or mandate to refine the formulae used by cellular network engineers to perform
cell mapping and base station distribution. The tendency instead has been to rely on relatively
crude cell-splitting techniques to accommodate increased traffic and, ironically, on ultrasophisticated
adaptive modulation and beam steering technologies for the same purpose,
neither of which, incidentally, require much engineering skill on the part of the network operator to execute. The first can be accomplished simply by building new base stations and
backing down on the power levels of those already in operation, and the second relies on the
intelligence programmed into the radio itself and not on the abilities of the network engineer.
However, a few specialized software tools exist for frequency reuse planning that have
recently become available for broadband wireless operators, and that you will consider in the
following section devoted to explaining the basic principles behind cell mapping.