Date: Thu, 3 Jun 1999 11:12:29 -0700
To: editor@telecom-digest.org
From: Arthur Ross 
Subject: Re: AT&T Now Charges For Place Name; Raises Other Charges


>[TELECOM Digest Editor's Note: It would not be a lot different, except
>perhaps for the immediacy of the circumstances if your neighbor's
>house caught on fire and he had never bothered to select some fire
>department to respond. Obviously you would have to call yours instead,
>and then attempt to get him to pay the bill for it later.
>
>I strongly and sincerely believe the *ONLY* legitimate function of
>government is to perform those services which the people being
>governed would find too inconvenient or impractical to perform for
>themselves. Protection against fires is one such task and other
>natural disasters. I would place most utility services in this same
>category. Arrangements for water, electrical and gas distribution
>on an individual basis would be very difficult or impossible. I am
>almost ready to say telephone networks should be in that category,
>but if telephones, then why not ISPs, the computer equivilent of
>the 'phone company'?  But as things are going now, it may not be
>long before ISPs are in fact regulated in much the same way telcos
>are regulated by federal and state agencies.   PAT]
>
Pat -

I tend to agree about ISPs, etc, but when it comes to the Ether, I think
it's more like water, sewer, etc. See attachments (3 pdf files total),
which I submitted to the FCC last fall in response to their request for
comments on "IMT-2000" spectrum policy. The point is that the new SSMA
wireless access techniques present some new,  creative ways to do this kind
of thing. They probably won't, of course. One of my FCC buddies told me
that even they had not thought this creatively (meaning the notion of
LEASING, not selling, spectrum, which is what they have been doing so far).

Of the 31 responses they got to this, 29 were from big corporations. The
other two were from me and one other guy, who works at another government
agency. I'm a citizen of the US, and as entitled as anyone else to have an
opinion. Figured, among other things, that it was a God-given
self-marketing opportunity, as every big-shot telecom law firm in the world
would probably read the responses.

Decided I must have been doing something right when, a few weeks later, I
got a large envelope from a big-shot DC FCC-type law firm, including some
sort of affadavit that "I have served this notice on the following ... "
and I was one of those parties named. What it was, was a complaint about
something Lucent had said in their response, but they had sent the
complaint to all of the respondees. Cool! Must have gotten their attention!

   -- Best
   -- Arthur

PS: Supposedly, this is available for public inspection at the FCC
somewhere, IAW their usual policies.

PPS: The attachment is text-only, which loses a little - there are some
small equations in the original pdf file that are lost here.


Maxwell, Shannon, and Economics
Spectrum Policy in the Era of Spectrally Efficient
Mass Market Wireless Communication
Dr. Arthur H. M. Ross
September 30, 1998


Introduction
As a participant in wireless standards activities, I welcome the FCC's
request for comments on IMT-2000 spectrum issues in preparation for the
upcoming World Radio Conference in the year 2000 (WRC-2000). The Commission
will doubtless receive numerous responses to this NOI from industry, both
operators and manufacturers. Those parties will naturally espouse US
positions that favor their individual businesses. I believe, however, that
the public interest will be best served by policy that takes collective
advantage of the unique technical properties of the emerging SSMA air
interfaces. This requires a new, possibly radical regulatory paradigm. I am
attempting to act as a public-interest advocate for that new paradigm.
Please note that the opinions expressed here do not necessarily represent
those of any of my clients, past or present. I would hope, however, that
widespread support will develop for these views, as they represent, in my
opinion, a win-win scenario for both the wireless industry and the
communicating public.
The decisions facing the Commission encompass several very difficult
technical problems and many imponderable policy issues. The most
appropriate course for the Commission is not only to recognize the
fundamental intractability and uncertainties of this multi-faceted problem,
but also to actually embrace them. A new regulatory regime founded on
flexibility rather than traditional carved-in-stone exclusive service
licensing, is now possible due to the new spectrally efficient spread
spectrum air interfaces. The primary regulatory consequence is that
licensing need not be exclusive. Multiple operators can be permitted use of
the same band in the same service area. Substantial benefits for vendors,
operators, and consumers ensue from such a regulatory regime.
The title of the paper stems from the three foundations of the wireless
business:
*	Principles of radio physics that make cellular wireless service
possible
*	Information theoretic concepts that govern communication performance
*	Business considerations
An enlightened synthesis of these three disciplines suggests the new
regulatory paradigm. The remainder of this paper discusses how such a
regulatory philosophy might work, why it is beneficial, and some of the
technical issues.
Current US Situation
At least one industry group in the United States has attempted to project a
requirement for IMT-2000 spectrum in the United States. The preliminary
conclusion seems to be that an additional 300 MHz is needed for mobile
(non-MSS) services. There is clearly a fundamental conflict between this
estimate and what is achievable, given the realities of global
electro-politics. There is almost no spectrum that does not have an
incumbent constituency that will vigorously oppose its re-allocation. The
opposition is often well-justified, particularly for applications like
radars and aeronautical aids to navigation that have safety-of-life
implications. Another problem exists in the Americas due to the PCS band
plan, which conflicts with the WRC '92 identification of 1885-2025 MHz and
2110-2220 MHz as the IMT-2000 spectrum. Not only is the Region 2 PCS band
plan different, it also overlaps the ITU-identified spectrum, so that
deployments of both plans in one geographical area is problematic.
While one can sympathize with the good intentions of trying to rationalize
a spectrum request for the WRC, the calculation that leads to the 300+ MHz
number is not credible and cannot be taken too seriously. It postulates a
demand number, a spectral efficiency per station, and some other system
characteristics. From those numbers it calculates a bandwidth, using a
straightforward formula. There are at least three problems:
=46irst, demand is not a number; it is a relationship. The usage of a servic=
e
will depend on the balance, as perceived by its consumers, between the
price of the service and the benefits they derive from it. The price that
the operator will charge depends on his costs, which certainly includes
fixed and recurring infrastructure costs, and potentially includes a cost
of spectrum, depending on the regulatory regime.
Second, the nature of the services is highly speculative. There are
IMT-2000 services projected for which there is no precedent, and thus no
credible market model.
Third, the spectral efficiency of the systems will be highly dependent upon
the nature of the traffic. The putative very high data rate packet
services, for example, will have very poor performance in low bandwidth
allocations. On average, the number of simultaneous users that can be
supported in a 5 MHz allocation at the highest data rate will be less than
one. The law-of-large-numbers averaging, which is the hallmark of spread
spectrum, is not effective when the number is not large.
In the best of all possible worlds, one might regard this as a very
complex, classical resource allocation problem, to which a sophisticated
linear programming model could be applied. Such a model would accept as
inputs the many factors that influence the business, and would output the
spectrum that should be allocated so as to optimize, in some sense, the
effects on US industry and consumers. However such a calculation is quite
intractable due to the many imponderables. These include market (read
"human") behavior, business models, political considerations, and
regulatory models. In the short term, we will have to make do with a
sub-optimal or alternative approach.
Guiding Policy Principles
Some guiding principles can be identified, based on the current rough and
tumble standards situation, and the difficulty of predicting future
technology trends. For the most part, these are already US policy, but they
perhaps bear explicit restatement.
*	Technology-neutral national policy - This policy is already in
place, and should be continued. The constraints on the operators should be
those of a fair and orderly marketplace, not an arbitrarily imposed
standard air interface. Although the ITU-R started with the goal of a
single worldwide IMT-2000 air interface, it is quite clear that this is
both undesirable and politically unachievable.
*	Flexibility
The regulatory regime should not dictate service offerings. Within broad
limits, the operators should be permitted to offer services over market
footprints that are dictated by its business considerations, not an
arbitrarily imposed service model.
*	Application-independence - The rules of the road should permit a
variety of applications, again for the operator to adjust service offerings
in accordance with marketplace realities. Things like out-of-band emissions
and maximum ERP should not be dependent on the nature of the service being
offered.
*	Traffic model independence - The nature of the traffic carried on
these future networks is difficult to anticipate. While one is tempted to
say that voice will be a large portion of traffic, even that may be carried
via an IP-like packet bearer. The regulatory regime should accommodate
widely varying traffic mixes, over both time and geography.
The overall regulatory goal should be a minimalist regulatory regime that
permits a maximum of flexibility on the part of the operators as to their
service offerings, and the air interfaces that they use. Adam Smith's
invisible hand will dictate the evolution of the business within the broad
regulatory constraints, whose primary purpose is to maintain a fair and
orderly marketplace.
Underlying Technical Principles
Land mobile radio is both facilitated and limited by the peculiar phenomena
that dominate near-earth propagation in the low GHz spectral bands.
Application of spectrally efficient principles of information theory permit
exploitation of this peculiar channel in near-optimal ways.
Maxwell - Radio Propagation in the Land Mobile Environment
The ugly communication channel that is the norm in the land mobile radio
environment is both a blessing and a curse.
It is a blessing in the sense that the natural and man-made surface features=
 result in larger-than-free-space path loss. The effective propagation law=
 is gene
rally modeled as a r__ propagation law, with _ ranging from about 3 to more
than 5, versus 2 in free space. Various explanations have been postulated,
but regardless of the physical reason, it is a well-established empirical
fact. It is a curiosity, in fact, that cellular radio would not work in an
infinite plane, uniformly populated with users. The integral that
represents the average interference power diverges like log(r). The
faster-than-free-space empirical law avoids this infinity.
It is a curse in the sense that the terrain and structures lead to
multipath propagation. Multipath propagation underlies the familiar
Rayleigh fading of the narrowband systems. It is particularly troublesome
to analog FM because it has direct adverse effects on the recovered audio.
The wideband CDMA systems are affected differently, but always adversely.
While the manifestations of the multipath differ in detail, the overall
result is a reduction in quality-of-service, or capacity, or both.
The multipath nature of the channel does have strong effects on the best
choice of modulation and spreading. Roughly speaking, the correlation time
of the signal should be of the same order as the correlation time of the
channel. While advocates of various products will claim superiority for
their particular choices, there really is no unique optimum answer because
there is no such thing as a typical channel. The nature of the channel is a
strong function of the physical environment. Dramatic differences are found
between indoor applications (both subscriber and base station indoors),
exterior urban, and exterior rural environments. What is optimum for one
environment may not be so for another environment.
Later we argue for large bandwidths as an aid to achieving large
aggregation of users. This must be done with due regard for the nature of
the channel, as excessively large bandwidth in a direct-sequence spread
signal will result in performance loss for a number of complex reasons.
This is not to say that it cannot be done, but only to say that it must be
done carefully.
Shannon - Fundamental Properties of the Noisy Communication Channel
Claude Shannon taught the world, fifty years ago, in 1948, that there is an
information rate, usually measured in bits per second, associated with a
communication channel, such that perfect communication can be achieved so
long as the actual rate does not exceed this channel capacity. For a
channel of bandwidth W, at fixed power, the channel capacity in bps is

where PS and PN are the signal and noise powers at the receiver. The
subsequent 50 years of development have seen the development of coding
schemes that achieve rates within a fraction of a dB of the Shannon limit.
This result holds for a power-limited channel with additive white Gaussian
noise.
Shannon further concluded that the optimum signals for communication in a
mutual interference limited situation are noise-like. While there was no
prescription given at the time of how to do this, subsequent developments
have produced approximations that  use digitally generated pseudo-noise
sequences. These can come quite close to the ideal noise-like model. Under
the assumption that all users arrive at the same power, and that there are
M>>1 of them, then the Shannon capacity rate for each is

or about  (ln 2)-1 =89 1.44 bits per second per Hertz of bandwidth, split
among the M users. Current systems actually achieve perhaps 10-15% of this,
in an environment that is much more complex and difficult than simple
additive white Gaussian noise. The lower figure includes all real effects,
such as handover complications real-world mutual interference, and power
control.
The Shannon rate corresponds to Eb/N0 =3D ln(2), or about -1.6 dB. In this
form it is more directly comparable to the performance metrics of realistic
systems. Current practice achieves Eb/N0 of the order of +3 dB to +12 dB,
depending on the channel, coding, and modulation. Moreover, the forward
link performance is a much more complex issue than this simple analysis
assumes, primarily because of handover considerations.
But even with the several dB deficit with respect to the Shannon limit, the
use of SSMA has improved spectral efficiency for the cellular voice
services by a factor of at least 5 or more.
Probably the most dramatic improvements in air link spectral efficiency
will come from clever uses of adaptive antenna technology. Use of the
so-called "turbo" codes will provide some additional improvements, as these
have been shown to come within a fraction of a dB of the Shannon bound in
additive white Gaussian noise.
Economics Meets Maxwell and Shannon -
How to best manage the scarce natural resource that is the electromagnetic
spectrum
Although some improvement may still be possible, dramatic new coding and
modulation techniques are not needed for greater spectral efficiency. The
simple benefits that accrue from SSMA have not yet been fully exploited.
=46ormerly we managed mutual interference by geographically sparse frequency
reuse. In cellular practice this is typically 21-way reuse, that is, each
channel assignment is used in only 1/21 of the stations (station meaning
one angular sector of one site). SSMA replaces this with universal
frequency reuse, but the mutual interference now appears in the Eb/N0
budget. Overall, a gain of perhaps 5-7 is realized, equivalent to a
frequency reuse of perhaps 3 to 4. The aggregation of interference
(pooling) that is achieved by SSMA is the key feature of the technique. The
law of large numbers has benefits for efficiency in the sense that the
variances of load and interference become fractionally smaller as the pool
of traffic is increased.
The law of large numbers is not effective for excessively large data rates,
as the number of users per station will eventually come to be not large.
This will be the case for all of the proposed air interfaces in the current
ITU-R RTT evaluation at the highest data rates, especially in the smaller
bandwidth channels. Variances in traffic, power, and other properties
become very large fractions of the averages when the number of users is
less than one.
Separation of Operators in Code Space
Current Commission rules allow any Commercial Mobile Radio Service (CMRS)
licensee the flexibility to change its radio transmission technology
without further approval beyond initial licensing. Only the most basic
technical compatibility constraints are imposed. In the cellular and PCS
services, a variety of air interfaces have in fact been chosen without
regulatory intervention. Each operator deploys its chosen technology within
its allocated frequency bands and geographic footprint. Coexistence is
enforced by simple, physical layer criteria that are often redundantly
imposed via both Commission regulations and air interface performance
specifications. Frequency reuse is at the discretion of each operator, but
only within his system.
The fact that even this partitioning works in practice is a consequence of
the near-earth, faster-than-free-space propagation.
=46or current purposes, the key observation is that this is an exclusive
licensing model, in the sense that each operator is granted an exclusive
license to use a particular duplex band pair, in a particular geographical
area.
But in the context of SSMA systems, frequency exclusivity is no longer
essential. In the same way that each operator can (and should) employ
universal frequency reuse for maximum system capacity, multiple operators
can make use of a common frequency band. Not only can users be separated by
spreading code rather than frequency, as we often like to say when
describing CDMA technology; operators can be separated as well.
Why share spectrum over operators?
In a word, aggregation. The primary benefits that ensue from the use of
SSMA arise from the aggregation of the mutual interference. It changes the
design criteria from worst case interference, manifested in the frequency
reuse patterns, to average interference. Aggregation of users within a
single operator's system helps; aggregation over multiple operators helps
more, due to the law of large numbers. In the case of some of the higher
rate IMT-2000 services, the estimated channel capacities are less than one
Erlang per station in a 5 MHz allocation. Aggregating load will lead to
greater air interface and trunking efficiencies. It may also the viability
of the business. It may be unprofitable for an operator to deploy high data
rate services if there is so little demand that his spectral efficiency is
poor. This could be a major influence in the decision to deploy the more
exotic services or not.
The difficulty, of course, is that sharing over operators complicates the
business and regulatory models. It is also not entirely obvious how it can
be done technically, although some possibilities are outlined below.
Benefits
Neglecting for the moment the technical aspects of how the sharing is
accomplished, one can identify some beneficial business consequences of
sharing spectrum. They are, in a sense, all economies of scale due to the
aggregation of traffic.
Resource Allocation
Carriers could fine-tune their support of spectrum in accordance with the
distribution over time and geography of demand for their particular
services. It would not be necessary to deploy support for the entire band
at all locations. Capacity could be utilized in greater or lesser amounts,
in accordance with demand.
Efficient Market
Amicable sharing agreements between operators could be implemented as a
sort of real-time auction, where the operators bid for the use of available
capacity, in accordance with their customer service agreements. An analogy
of this might be found in the real-time trading of electric power that
takes place among utilities in accordance with the shifting of supply and
demand over geography and time.
Creative Service Offerings
Service pricing could be in accordance with the very real tradeoff between
supply and demand. For example, low prices could be attached to services
that utilize extra capacity during off-peak hours, or to use of business
services used in residential neighborhoods, or vice versa. Conversely, high
data rate premium services could be supported for those customers who
demand it, with the extra cost of providing that service passed through in
the form of higher prices.
Kinds of Spectrum
Several different classes of licensed spectrum can be envisioned,
distinguished by their sharing status. They are:
Exclusive use, by frequency and geography - This is the current licensing
model. The licensee gets exclusive use in the authorized band and market.
It is at the discretion of the operator what air interface and services he
offers. In the current US usage, these are all frequency duplex, with the
duplex spacing specified by the license.
Multiple use, by frequency and geography - Two or more operators get use of
the authorized band and market. It is up to those operators how they
implement the sharing and the uses to which they put its capacity.
Uncommitted Pool - An uncommitted band, over a particular geographical
area. Capacity could be purchased, as needed, in particular areas. That
purchase could be by advertised price, or by real-time auction.
It could be anticipated that most operators would opt for a partially
exclusive license, to support their core services, with a shared component
to support other, more competitive, variable services.
Duplexing in the shared-use options is a question for debate. One
possibility is that there might be two (or more) sub-categories: one with
the same, fixed duplex spacing as the exclusive licenses, and another
category where uplinks and downlinks are managed separately. Only the most
basic technical regulatory constraints are imposed on the licensees, aimed
only at other-service interference control, and health and safety issues.
Asymmetric allocations are often mentioned, under the assumption that human
typing data rates are much lower that video screen painting rates.
Implementation of Shared Allocations
The non-exclusive shared licenses could be managed in several ways. Some
examples follow.
Dynamic Footprint (Negotiated Frequency-Geography Partitioning)
A frequency-time partitioning could be negotiated among the shared
licensees. Functionally it would be similar to the exclusive license, but
done privately rather than by regulatory action. It also might change with
time, subject to a real-time, automated negotiation. Any particular band,
in any particular station, would be used by only one operator at a time.
This presents a problem of what to do at the dynamic boundaries between
operators. The simplest procedure, that of guard bands and guard zones is
also the most wasteful. Other procedures, involving multiple support at the
border cells can be envisioned, but it a technical challenge. For example,
in IS-95-like systems, the border cells could be populated with pilot
beacons, which would be used to trigger a frequency change for those
subscriber stations that need to handover across the boundary. Those
subscribers would be moved to the exclusive portion of his operator's band
before executing the handover.
Non-orthogonal Overlay
The sharing participants could negotiate a technical overlay. Two different
air interfaces would be supported through a common antenna in the same
band. They must cooperate on issues of power control so as to properly
share the available channel capacity. The capacity used by each could be
either held constant, or could itself be negotiated real-time in accordance
with a pre-arrangement.
There is a performance penalty for the non-orthogonal overlay, as the
incompatible signal manifolds mutually interfere as radiated from the base
station. Orthogonal overlay or a virtual system is preferable.
Orthogonal Overlay
An operating model in which orthogonality is imposed in the manifold of
signals radiated from one station, but they are otherwise logically
independent. There is no proposal corresponding to this kind of concept.
The virtual system is probably the optimal variant of this, although it
poses regulatory and business model questions.
Virtual System
A virtual system is one in which a single, high capacity, common downlink
is used to support multiple operators. The uplink could be either the same
air interface protocol, or something completely independent, as the level
mutual interference of the uplinks is not normally dependent on what air
interface they use. The appearance of separate systems is synthesized via
the network software. The network architecture, that is, whether traffic is
separated by the base stations themselves, or it is carried out at a
central facility (BSC or MSC) is an implementation issue.
This is the form of sharing that carries the least risk.
Kinds of Base Stations
Existing systems have been deployed largely independently, although in some
cases antenna towers are shared. In a non-exclusive license scenario,
however, there are strong incentives to collocate facilities, and for
several reasons.
Non-collocated Base Stations
This is the current model in most cellular and PCS installations. It must
be restricted to exclusive band plan operation because the near-far problem
would created intolerable mutual interference in a common band. Their base
stations and subscribers would severely jam one another in cross-system
proximity situations.
Non-collocated base stations can use nominally exclusive bands
simultaneously, but they will have to observe guard zones and guard bands
in order to avoid incidental interference.
Collocated Base Stations
Collocated stations are technologically very helpful in all sharing schemes.
Collocation helps the out-of-band emissions problem in exclusive band plan
operation because the spurious emissions will have a fixed relationship to
the desired signal. The interference level is set at the transmit antenna
and is insensitive to subscriber set location.
Collocation is required for all frequency overlay scenarios because of the
power control issue. Achievement of maximum channel capacity requires that
all in-band signals, desired and undesired, arrive at the receiver at the
minimum necessary power. This cannot be achieved without collocation.
Receiver Intermodulation Problem
It has been found in the early deployments of CDMA systems that serious
interference problems occurred in CDMA handsets when they were too close to
AMPS stations operating in an adjacent band. This is caused by nonlinear
mixing products of the AMPS carriers, generated within the receiver front
end, that fall inside the IF  passband. This is a problem that will always
exist at some level: there always will be a "kill zone" around alien base
stations due to such intermodulation products. The only question is the
size of that zone. While collocation is not required in exclusive band plan
situations, doing so will largely mitigate the co-channel intermodulation
problem.
Most historical experience with intermodulation problems has been in
AMPS-CDMA adjacent channel situations. However the problem is not
restricted to narrowband-on-CDMA. It does not seem to have been encountered
much by PCS operators yet, but as the D, E, and F block operators begin to
turn on, CDMA-on-CDMA intermods can be expected to become an issue as well.
Secondary Benefits
The sharing of RF transceivers and antenna structures, while primarily
motivated by considerations of communication performance, also has
favorable side effects from the standpoint of public policy. Sharing
facilities reduces the costs, over all operators, of the inevitable legal
wrangling over local zoning, environmental issues, real estate, and
maintenance costs. This is a savings in the cost of doing business, and
should ultimately benefit the end consumer.
Global Control Channel
The recent discussions in the ITU over the global control channel, an
agendum for WRC-2000, have missed the point. A global control channel of
some kind is needed in order to carry out the dynamic spectrum management
that we have outlined. It would broadcast the band plan, which might change
over time and geography in accordance with the sharing agreements and
procedures. It might serve, in fact, as a truth-in-advertising vehicle,
i.e. it could offer a shopping list of services and billing rates from
which the subscriber set could select the purveyor of choice, based on
subscriber-defined service choices and price criteria.
What is the Ether1 Worth?
While this is obviously a politically charged issue, a regulatory regime
that leases spectrum, with use-based payments would appear to present an
attractive alternative to the one-time auction process. But how would this
work in the context of station-by-station, hour-by-hour changes in band
configuration and load? What constitutes "use" of the electromagnetic
spectrum for billing purposes?
Clearly "consumption" of ethereal capacity by communication links is
related to the energy radiated and its geographical distribution. The
radiation from each transmitter degrades the noise figure of all
non-targeted receivers to some degree. The deleterious effect of a
transmitter can be viewed as a form of receiver heating within its antenna
coverage footprint. To the extent that the interference is equivalent to
Gaussian noise over the signal bandwidth, the apparent noise figure of
impacted receivers will rise due to the interference. The fee might be
viewed as a sort of pollution charge based on the net degradation to those
other potential users.
A plausible metric might be a time integral of radiant power flux, averaged
somehow over the non-targeted receivers that are adversely impacted by the
radiation. In practice this probably would take the form of a negotiated
weighting factor derived by integrating the antenna pattern over a
demography map. If all the radiation can be concentrated at the intended
receiver, and no others, the adverse impact, and hence fee, is zero. The
incentive is to implement the system with the minimum possible energy
needed to communicate. And the metric is indeed energy, not power. A large
power radiated for a short time is treated as equivalent to a small power
radiated for a long time if their time integrals are equal. For any
particular combination of modulation and coding, error rate performance can
be characterized by the signal-to-noise ratio Eb/N0.
Business model
The notion of shared spectrum licenses, while attractive technically for
the reasons already cited, presents some interesting business model
questions. Technical considerations very much favor implementation of
shared or community facilities, encompassing primarily the base station
transceivers and antennas. There are also economies of scale and favorable
effects on zoning boards and local jurisdiction site licensing. However,
from the standpoint of regulatory policy this would appear to be a step
backwards, contrary to the trend toward deregulated utilities,
transportation, and communication companies. Who might own and operate such
shared facilities? Would it be a regulated, monopolistic entity akin to the
old Bell System? Would it be a joint venture of  the local operators? Or
would it be a wholesale operator who would own and operate the radio
infrastructure. The wholesaler would resell capacity to other business
entities, who would be the visible purveyor of wireless services to the
consumers. Whether the wholesaler owns the network too, or stops at the
base stations, is open to debate.2
It is possible here that there might be a role for some central, regulated
monopolistic entity to operate what might be regarded as a national
wireless information utility. This would be, obviously, a very
controversial proposal in this era of world wide deregulation and
privatization of government owned utilities. There are, however, strong
analogies with other kinds of persistent utilities like water, sewer,
postal service, electric power. Clearly any such proposal would be the
subject of a spirited public debate.
Recommendations
The Commission, in its infinite wisdom, would do well to explore this
concept in some detail, consulting with the expert communities in
communication technology, public utility policy, business, and economics. I
suggest the following concrete objectives for such consultation and
rulemaking:
1.	Determine the viability of such a creative new regulatory regime
permitting shared spectrum licenses. This might include a study of the
viability of the broadly tunable mass market subscriber sets that would be
needed to support such a regime.
2.	Determine the minimum set of technical issues that would need
regulation.
3.	Draft strawman regulatory policy and licensing procedures.
4.	Search the current United States spectrum allocations for candidate
bands suitable for a reassignment to this new service, with a preference
for bands below 3 GHz. One of the primary purposes in suggesting this
flexible policy is to permit a degree of ad hoc spectrum management, so
these candidates need not necessarily align with the ITU identifications.
This is particularly desirable given the current disconnect between ITU
Region 2 and the rest of the world in the PCS bands.
5.	Initiate a formal inquiry in accordance with Commission procedures.
1 Ether: The pervasive, infinitely elastic, massless, medium postulated,
erroneously, by 19th century physicists, as the medium of propagation of
electromagnetic radiation. The term is used here as a picturesque
placeholder for the asset value of electromagnetic spectrum.
2 It is perhaps noteworthy that this is more or less the business model
that was attempted by Nextwave Telecom in their C-block PCS system. While
this failed for other business reasons, the concept is technically sound.
Report Title

Section Title

Page 2	Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited. All rights
reserved.

Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited		Page 1
All rights reserved



Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited. All rights reserved.

Maxwell, Shannon, and Economics



Page 10	Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited. All rights
reserved.

Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited. All rights reserved
	Page 9



Copyright (c) 1998 Arthur H. M. Ross, Ph.D., Limited. All rights reserved.

--============_-1283688351==_============
Content-Type: text/plain; charset="us-ascii"

   -- Dr. Arthur Ross
      2325 East Orangewood Avenue
      Phoenix, AZ 85020-4730
      Phone: 602-371-9708
      Fax  : 602-336-7074
--============_-1283688351==_============--