“New Approaches to Licensing”
Oliver Holland, King’s College London
Arturo Basaure, Aalto University
Numerous other authors, e.g., in SIG-PL, etc.
Workshop on Cognitive Radio Policy and Regulation
Vilnius, Lithuania, April 2014
Structure
 Introduction
 New Approaches to Licensing
 Light Licensing
 Authorised Shared Access
 Licensed Shared Access
 Other Developments
 Our Proposal: Pluralistic Licensing
 Conclusion
Introduction
The Past: License Exempt
 Simple basic rules to access of spectrum freely, such as
transmission power limit, spectrum mask
requirements, etc. Assessed at device certification.
 Pros:
 A hot-bed of innovation in wireless field—created a plethora of




novel applications and wireless uses.
An entry point for “free” wireless services.
Led to the development of popular standards such as WiFi,
Bluetooth and ZigBee.
Spill-over of related innovation to other bands.
Rich eco-system.
 Cons:
 Interference/coexistence concerns, coverage issues, others.
The Past: Licensed
 Users have to obtain a license to access the spectrum, again
with some rules attached by the regulator. Devices again
assessed at certification, as well as means being enforced
for ensuring that license is adhered to.
 Pros:
 Reliable services:




Near-ubiquitous voice connectivity in public land mobile.
Early data services.
High-quality broadcast services.
Guaranteed clean spectrum for critical services.
 Cons:
 There are efficiency concerns of such bands .

Parallel need for more spectrum, mobile data growth, etc.
The Present: Stalled Innovation?
 Opportunistic/dynamic spectrum access (OSA/DSA), enabled by
cognitive radio key impetus of wireless research & innovation.
 Barriers:
 Extremely onerous technical requirements to protect incumbents.
 Nascent markets.
 Geolocation database access and location determination challenges.
 Lack of appropriate spectrum.
 Resistance of incumbents due to “fairness factor”.
 Severe and yet to be defined certification requirements.
 Difficulty to scale from test & trial licensing incentives (low cost) to
commercial deployment (extremely high cost):

Costs – e.g. nationwide, long-term licence.

Deployment band & conditions of usage are usually different than the wireless test
& trial scheme used to test the new technology/service.
Moving into the Future: Data Growth
 The mobile/wireless telecommunication market expected
to continue growing extremely rapidly in the next decade.
 Presents many challenges:
 Provision of spectrum.

Both a problem and solution.
 QoS guarantee for the given data rate.
 Energy efficiency.
 Sufficient (quality) spectrum for new players in the market—
competition concerns.
 Numerous others…
Moving into the Future: Data Growth
M2M
Moving into the Future: Data Growth
 Cooper’s Law, Edholm’s Law, etc.
illustration
courtesy of
IEEE Spectrum
Moving into the Future: Data Growth
 Capacity provision. Solution is to increase:
 Rate per bandwidth (radio interface performance / spectral efficiency).
 Overall bandwidth (amount of spectrum).
 Densification (frequency reuse).
 Shannon limit challenging, and already being reached (if systems are tested in
idealistic scenarios). “Densification” presents severe challenge (among others)
of there being a vast increase in access points—latent energy consumption,
securing of sites, management issues, etc.
 Increase of spectrum is an appealing solution to address such issues – can be
done (usually) with the same sites, (in some cases) similar equipment, lesser
implications in terms of energy consumption, etc.
Need for new Licensing Schemes
 Spectrum sharing is key means for provision of more spectrum.
 Many services do not use all of their spectrum.
 In time.
 In space.
 In a spectral sense, e.g., channels within a band allocation.
 Technical means exist that enable the reuse of spectrum.
 Temporally.
 Spatially.
 Spectrally (e.g., reuse of channels).
 Appropriate safeguards can usually be implemented in allowing sharing.
 Regulation is the biggest challenge for spectrum sharing technologies.
 Can be extrapolated to licenses as issued by regulators not allowing
spectrum sharing and related technologies such as CR. Licenses that
permit reuse of unused spectrum of an incumbent in a controlled
manner, guaranteeing no harmful interference to the incumbent, needed.
New
Approaches to
Licensing
Light Licensing
 (Sort of) new. Has long been the case that spectrum is
allocated (though different priorities) on a spatially varying
or individual basis. But novel light licensing models are big
advancement, with appropriate safeguards, accelerating
regulator-authorised sharing among (co-)primary systems.
 Such solutions can nevertheless often still imply a workload
for the regulator or associated entity, due to the process of
allocating the license, calculating implications, etc.
 One model (for infrastructure-based scenario).
 Small fee for non-exclusive use of the spectrum in general.
 Additional license fee for each deployed base station.
 Client devices are not required to pay a fee, but must receive an
enabling signal from a licensed base station before transmitting.
Light Licensing
 Also subject to alternative understandings/implementations,
anywhere between the following extremes.
 One extreme:
 Individual authorisation (e.g., per-BS deployment is example):
 Individual frequency planning/coordination.
 Simplified procedure compared with typical procedure of individual
authorisation.
 Likely limitations on the number of users.
 Other extreme:
 A form of general authorisation characterized by:


No individual frequency planning/coordination.
A requirement for registration and/or notification that allows
controlling the deployment and use of the application, but does not
restrict it.
Light Licensing
 Being suggested for deployment (or being deployed) in
various prominent contexts. Just a few examples here.
 IEEE 802.11y:
 High powered data transfer equipment to operate on a co-
primary basis in the 3.65 to 3.7 GHz band in the United
States.
 Other potential bands 4.9-5GHz, even IMT-Advanced
bands.
 Other services (e.g., WiMAX) in 3.65 to 3.7 GHz band.
 E-band 71-76 and 81-86 GHz, “pencil beam”
communications.
Light Licensing
 “Second digital dividend” in ITU region 1.
 Mobile broadband and broadcasting on co-primary basis
in 694MHz-790MHz (lower bound and general rules to be
determined in WRC 2015).
 Could be candidate for light licensing type agreements?





Will be clearer after WRC 2015, and after result national
regulatory decisions.
FCC’s 3.5 GHz small cells (“General Authorized Access”).
5.8 GHz spectrum to allow other services coexistence.
Professional PMSE equipment.
Amateur services.
 E.g., Ofcom business radio licensing.
Authorised Shared Access (ASA)
 Scheme proposed by Qualcomm and Nokia in a joint response to an
RSPG consultation on cognitive technologies in January 2011.
 Subsequently refined in report and presentation to ECC meeting in
March 2011, and in a May 2011 in a report to the CEPT Working Group on
Frequency Management.
 Based on the use of cognitive radio techniques to determine channel
availability.
 Bilateral negotiated and regulator recognized agreements between new
and incumbent users to set the conditions for frequency access.
 Might include compensation to the incumbent(s) for agreeing to share.
 Newcomer, in return, would gain an assured amount of spectrum
availability in the form of a guaranteed minimum amount of spectrum
use time in certain geographic areas; advance warning if access will be
suspended.
See, e.g., http://www.qualcomm.com/media/documents/files/wireless-networks-1000x-more-spectrum-especially-for-small-cells.pdf
Authorised Shared Access (ASA)
 Proposed in context of reuse by mobile systems, but not limited to that.
Image courtesy of Qualcomm
Licensed Shared Access (LSA)
 As defined by the Radio Spectrum Policy Group:
“An individual licensed regime of a limited number of licensees in a
frequency band, already allocated to one or more incumbent users, for
which the additional users are allowed to use the spectrum (or part of the
spectrum) in accordance with sharing rules included in the rights of use of
spectrum granted to the licensees, thereby allowing all the licensees to
provide a certain level of QoS.”
 Translated (!!):
 Framework for limited number of additional licensees to obtain access
to spectrum that is already allocated. Licensed users (“incumbents”)
share spectrum with one or more new users (“incoming users”) in
accordance with conditions defined by regulation (based on
agreement of both “licensed” and “opportunistic” parties).
 These conditions may be “static” (e.g., an exclusion zone, with a
restricted time for operation) or more “dynamic” (e.g., geographic /
time sharing, on-demand authorisation).
LSA vs. ASA
 Distinction between ASA and LSA not widely defined, even often (in a





number of high profile publications) taken to be one and the same thing.
LSA is at least an attempt to codify ASA at the regulatory/policymakers
level, and debatably somewhat more than that.
LSA brings licensing up-front in that it makes it clear that the
opportunistic LSA/ASA spectrum user is also issued a license, etc.
LSA takes negotiated agreements concept between both parties on
guaranteed minimum QoS.
But considers any new agreement on temporary transfer of frequency use
rights as a “change of use” for an exclusive assignment. Therefore, has the
“sharing rules included in the rights of use of spectrum granted to the
licensees”. I.e., the “sharing rules” negotiated by the licensees must be
approved by the regulator and incorporated in their licence conditions as
amendments or replacement licenses.
LSA seems to be moving away from incorporation of “cognitive radio ”
approaches in the spectrum reuse by the ASA licensee.
Other Developments
 There are numerous other developments here. Pertinent examples:
 White space rules per se, although these generally are expressed as appropriate
conditions/rules for the “unlicensing” of white space equipment1, 2.
 FCC “small cell” three-tiered access in 3.5GHz. Resonances with ASA/LSA,
light licensing approaches, and white spaces database approaches 3.
 FCC “interference limits policy” approach. Means to greatly enhance the
efficiency of spectrum access, given knowledge of implied interference at the
receiver, rather than transmitter-side (e.g., power limit)4.



1
2
3
4
Licenses instead based on interference limits…?!
Second incarnation of this concept, after initial suggestion some 10 years ago.
Associated efforts to characterise receiver-side, e.g., rejection of adjacent band
interference, etc.
FCC Second Memorandum and Order, “, Additional Spectrum for Unlicensed Devices Below 900 MHz and in the 3
GHz Band” September 23, 2010 (see also the third MO&O from August 2012!!)
Ofcom, “A Consultation on White Space Devices Requirements,” (and related consumtations linked to therein),
http://stakeholders.ofcom.org.uk/consultations/whitespaces
FCC NPRM and Order, “Enabling Innovative Small Cell Use In 3.5 GHZ Band NPRM & Order,” March 2013
FCC Public Notice:, “Office of Engineering and Technology Invites Comments on Technological Advisory Council
(TAC) While Paper and Recommendations for Improving Receiver Performance,” June 2013
Our Proposal:
Pluralistic
Licensing
Pluralistic Licensing
 “Award of licenses under the assumption that opportunistic secondary
spectrum access will be allowed, and that interference may be caused to
the primary with parameters and rules that are known to the primary at
the point of obtaining the license”.
 Primary will choose from a range of offered “pluralistic licenses” each with
associated fees, and each specifying alternative opportunistic secondary
spectrum access rules with known interference characteristics.
 Objective is to incentivize the primary to obtain this type of license
through means such as a reduced license fee, whereby the opportunistic
secondary spectrum access will use “cognitive radio” mechanisms to avoid
causing interference to the primary or to otherwise keep interference
within known parameters.
 Better secondary design can also be incentivised, e.g., through reduced
certification fee (however, not a core objective of the concept).
 Novel means for spectrum licensing which is fair to both primary and
secondary users and takes into account requirements of both parties.
Pluralistic Licensing
 Why not combine best of the licensed/licence-exempt worlds?
Licensed access:
Opportunistic access:
Reduced fee compared
with conventional
license; accepts
opportunistic access
Low or zero fee; must
meet requirements for
awareness of primary
users
Pluralistic
licensing
See, e.g., O. Holland, L. De Nardis, K. Nolan, A. Medeisis, P. Anker, L. Minervini, F.
Velez, M. Matinmikko, J. Sydor, "Pluralistic Licensing," IEEE DySPAN 2012, Bellevue,
WA, USA, October 2012
Pluralistic Licensing
 Benefits:
 Encourages more robust or better design of primary systems, leading
to the ability to cope with increased degree/risk of interference among
other benefits.


Greatly improves spectrum usage efficiency and fairness.
Possible pricing mechanisms for secondary systems (e.g., at certification stage?)
might also lead to the better design of secondary systems—e.g., incorporating
sensing for secondary-secondary awareness (better secondary coexistence).
 Allows opportunistic spectrum use while providing any level of
protection desired by the primary.
 Highly flexible dependent on the case-by-case deployment context
(e.g., intended primary service), even so far as allowing solutions such
as spectrum sensing and primary beacon transmissions which would
otherwise not be practical or desirable, while defaulting to safe
mechanisms such as a geolocation database in cases where a lower
interference variance is required.
Pluralistic Licensing
 Benefits:
 Reduced challenge for secondary coping with inefficiencies of legacy
primary systems (particularly in “green-field” deployment scenarios).

In obtaining the license, primary users implicitly accept rules of the band hence
will be designed and manufactured with better technical capabilities that are
able to cope with those rules—better rejection of adjacent channel interference,
better robustness to short-lived interference increases, better sensitivity, etc.
 Highly scalable to progressive deployment in more spectrum bands.
 Rules for each deployed band (e.g., the use of spectrum sensing, the geolocation
database, or potentially even a beaconing mechanism being employed by the
primary) being determined based on the intended context for that band (e.g.,
the expected primary services(s), expected burden on the primary, etc.).
 Lends well to progressive deployment of primary transmitters as a network
invests, for example, using a licensing regime similar to light-licensing.
 Could even be (carefully) deployed in legacy bands.
 Incentivises the primary to make more efficient use of spectrum, and
to make unused spectrum available.
Pluralistic Licensing
 A good, adaptable balance between exclusive use and license-exempt
access.
Licenceexempt
Lightlicensing
Pluralistic
licensing
more
interference
Registration with
e.g. ISM,
short range regulator, e.g., low
power in-building
devices
system
less
interference
Authorised /
licensed
shared
Exclusive
access
e.g., mobile
broadband/
2.3GHz
e.g.,
GSM/UMTS/
HSPA/LTE
 Primary interference and license fee based on primary/secondary
services.
Interference
Dual-priority broadband
License fee
Primary broadcast; secondary
mobile broadband services
Pluralistic Licensing
 Proactive primary mechanisms and pricing.
Burden on primary
License fee
Beacon messages sent by
primary to ward off secondary
Primary registered in
geolocation database only
 Possible pricing example for secondary (relevant to secondary-secondary
coexistence; could be relevant also to primary-secondary).
Burden on secondary
High quality sensing
(e.g., feature detection),
for secondary politeness
Fee paid by secondary
Low quality sensing
(e.g., energy detection)
No sensing (only refers
to geolocation database)
Pluralistic Licensing
 Example considerations in choosing/configuring PL.
 Fixed vs. mobile (primary and secondary)
 Strongly affects interference dynamic need to monitor/adapt, etc.
 Packet-data or not (primary)
 Packet-data usually more robust.
 Simplex vs. duplex (primary)
 Duplex usually implies feedback channel: ability to adapt to
changing situation in terms of interference or channel quality.
 Analogue vs. digital (primary).
 Increasingly rarely a concern nowadays, but usually digital is more
robust.
 Selection of, or scope for variation of base–level primary
characteristics to mitigate interference, e.g., even simplex
could be more robust if the coding rate were higher.
Assessment of the Concept
 COST-TERRA STSMs (Arturo Basaure, Aalto
University, visiting KCL in May 2013 and April
2014.
 Analytically and through simulation assess the
effects of changes in rules (e.g., TV white space) on
the primary, and map to means for compensating
the primary for the change in performance,
investigate rules for different scenarios (e.g.,
different types of primary/secondary systems).
 Simulation system been created, and
parameters/scenarios changed to gauge effects.
Assessment of the Concept
 Primary users: LTE-like FDD
network
 Secondary users: P2P
communication
 Performance for primary
users connected to the
”central” (pink) base station
in the above assessed
 Performance of secondary
users and achievement for the
spectrum as a whole also
assessed
Assessment of the Concept
 FCC Rules: Effects on SINRs and total number of supported users of
changing the distance moved after which secondaries must check again
with the geolocation database whether they are allowed to transmit
Total number of users
(e.g., useful for vehicular scenarios). 40
35
30
25
20
100m
15
200m
300m
SINR ( average values)
10
25
5
23
0
40
21
primary SINR
(100m)
secondary SINR
(100m)
primary SINR
(200m)
secondary SINR
(200 m)
primary SINR
(300m)
secondary SINR
(300m)
19
17
15
13
11
9
7
5
40
43
47
50
Primary base station power (dBm)
53
43
47
dBm
50
53
Primary base station power (dBm)
Assessment of the Concept - Profits
 Utility functions: mapping SINR to value
p(SINR)
User utility functions
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
real time
not real
time
strict
real time
0
10
20
SINR (dBm)
30
Assessment of the Concept - Profits
B. Not real-time primary / real-time secondary *
A. Real-time primary / not real-time secondary *
Profits (USD)
Profits (USD)
500
500
450
Rp
400
450
400
350
Rs
300
350
300
Rp + Rs - C
250
200
C
150
250
200
150
100
Rp + Rs
50
100
50
0
40
43
47
50
BS transmission power (dBm)
0
53
40
C. Real-time primary and secondary *
43
47
50
BS transmission power (dBm)
53
D. Strict real-time primary / not real-time secondary **
Profits (USD)
400
Profits (USD)
600
350
500
300
Rp
250
Rs
400
200
Rp + Rs - C
300
primary
profits
secondary
profits
C
150
Rp + Rs
100
200
total profits
100
50
0
0
40
43
47
50
BS transmission power (dBm)
53
100
200
leasing distance (m)
300
Assessment of the Concept - Profits
D. Strict real-time primary / not real-time secondary **
Profits (USD)
600
Increase in value of spectrum
500
primary
profits
400
secondary
profits
300
200
total profits
Secondary ~150 USD gain –
can more than compensate
primary and still make very
healthy profit
100
0
100
200
leasing distance (m)
300
Primary ~70 USD loss
What’s New? –And Implications
 Primary selecting from many possible licenses as it would
require, embodied in a well-defined/automated mechanism.
Incentive scaling to primary based on implied interference.
 LSA implies that “secondary” is licensed, Pluralistic Licensing
implies no license for secondary – simpler/quicker to
implement.
 Although there are numerous definitions/understandings of
ASA/LSA, the direction seems to be toward LSA not including
“cognitive” approaches—PL opportunistic access is as flexible as
primary would like so does include option of such approaches.
 Could lead to very spatially/temporally dynamic primary
license situation – should be twinned with database that
secondary will check for rules in area. This could be
extension of TVWS geolocation database.
Our Suggestion
 We argue that “pluralistic licensing” should
become the de facto form of license awarded in
newly designated bands.
 Should also be offered to the primary as a licensing
option in existing bands in cases where
interference to legacy devices that might exist in
the band can be avoided.
 Under such an approach, opportunistic
secondary spectrum access can be realised
today.
Conclusion
Conclusion
 Increased spectrum sharing is necessary, e.g., to satisfy




future capacity demands.
Current licensing schemes cannot sufficiently realise
spectrum sharing and associated facilitating technologies.
 There is a need for novel licensing schemes.
There is considerable progress to these ends but further
work needed in a number of areas.
Have presented some of the key developments here, also
our proposal for “Pluralistic Licensing”.
Numerous concerns must be addressed, e.g., fairness to
incumbents of spectrum sharing (multi-faceted problem),
motivations for existing and new stakeholders, etc.
That’s all folks!
oliver.holland@kcl.ac.uk
www.ollyinteractive.com
Acknowledgments
 This work has been prepared in the framework of and
supported by COST IC0905 “TERRA”, www.cost-terra.org,
and has also been supported by the ICT-ACROPOLIS
Network of Excellence, FP7 project number 257626,
www.ict-acropolis.eu, and ICT-SOLDER, FP7 project
number 619687, www.ict-solder.eu
 Many thanks to other contributors to the Pluralistic
Licensing concept: Keith Nolan, Luca De Nardis, Arturas
Medeisis, Peter Anker , Leo Fulvio Minervini, John Sydor,
Marja Matinmikko, Fernando Velez, Vania Goncalves, Raul
Chavez-Santiago (list not exhaustive)