SPECTRUM MARKETSusers.ece.northwestern.edu/~mh/MSIT/spec_markets.pdfspectrum sharing across military...
Transcript of SPECTRUM MARKETSusers.ece.northwestern.edu/~mh/MSIT/spec_markets.pdfspectrum sharing across military...
-
SPECTRUM MARKETS
MSIT Week 10
Michael Honig Department of EECS
Northwestern University
March 2014
-
Spectrum Markets
Policy Economics Engineering
Rakesh Vohra, Kellogg MEDS
Randall Berry, Michael Honig, EECS
2
-
High Profile Issue 3
-
Why Spectrum Markets?
¨ New policies are needed for spectrum allocation. ¤ Markets are natural policy candidates.
¨ Markets for spectrum pose unique challenges/questions. ¤ Definition of property rights, interference externalities ¤ Efficiency, incentives, wireless system design
¨ Interplay between economics and engineering issues
4
-
Background Spectrum Sharing Models Market structures How should spectrum be managed?
Outline 5
-
Limited Supply of Spectrum
good for cellular (300 MHz to 3 GHz)
6
-
Increasing Demand
Spark-gap transmitter
(Tesla, 1893)
7
-
Spectrum “Crunch” P
etab
ytes
per
mon
th
8
-
Regulation Prior to 1927: Open to All
Earliest uses of wireless for ship-to-ship, ship-to-shore communications.
Broadcast radio begins in 1921. Licenses issued by the Department of Commerce.
9
-
Two Landmark Cases
Hoover vs Intercity Radio, 1923 United States vs Zenith Radio, 1926
Herbert Hoover, US Sec. of Commerce
Department of Commerce has no authority to regulate licenses.
è Broadcasting boom: 200 new stations appeared in < 6 months. Interference created “chaotic” radio environment (tragedy of the commons).
10
-
Spectrum Property Rights: A False Start
¨ Congress subsequently passed legislation prohibiting spectrum property rights
¨ Licenses issued for 90 days.
¨ Tribune vs Oak Leaves Broadcasting, 1926 ¤ Property right allowed based
on “homesteading” ¤ Interfering stations could be fined.
11
-
Regulation since 1927: “Command and Control”
¨ Federal Radio Commission (FRC) established in 1927.
¨ Federal Communications Commission (FCC)
established in 1934.
¨ Maintains authority to ¤ Grant / renew / deny licenses for spectrum use. ¤ Assign applications to particular frequencies. ¤ Police content and use
“Wise old man approach to spectrum allocation”
12
-
The Spectrum Paradox
¨ Spectrum is a scarce resource ¨ Spectrum is underutilized
13
-
Spectrum is a Scarce Resource
“beachfront property” Nearly $20B netted for 700 MHz auctions in 2008.
14
-
Spectrum is Underutilized
Spectrum measurements in New York City and Chicago conducted by Shared Spectrum Co.
15
-
An Economist’s Proposal
Ronald Coase, 1991 Nobel Laureate in Economics
Introduce spectrum property rights, sell to highest bidders, do not restrict use.
R. Coase, “The federal communications commission,” J. Law and Economics, pp. 1–40, 1959.
Coase’s “Theorem”: In the absence of transaction costs, spectrum owners will trade rights so that the outcome allocates spectrum to best use.
16
-
An Economist’s Proposal
Ronald Coase, 1991 Nobel Laureate in Economics
R. Coase, “The federal communications commission,” J. Law and Economics, pp. 1–40, 1959.
Coase’s “Theorem”: In the absence of transaction costs, spectrum owners will trade rights so that the outcome allocates spectrum to best use.
Role of government should be to minimize transaction costs.
Introduce spectrum property rights, sell to highest bidders, do not restrict use.
17
-
An Economist’s Proposal
Ronald Coase, 1991 Nobel Laureate in Economics
R. Coase, “The federal communications commission,” J. Law and Economics, pp. 1–40, 1959.
Spectrum auctions finally introduced in the 1990s. Restrictions on use remain.
Coase’s “Theorem”: In the absence of transaction costs, spectrum owners will trade rights so that the outcome allocates spectrum to best use.
Introduce spectrum property rights, sell to highest bidders, do not restrict use.
18
-
A Governor’s Proposal
Rod Blagojevich Former governor of Illinois
Introduce property rights for senate seat, sell to highest bidder.
19
-
A Governor’s Proposal
Rod Blagojevich Former governor of Illinois
Introduce property rights for senate seat, sell to highest bidder.
Found guilty of violating federal laws…
20
-
Exclusive use (liberal licenses) Commons Hierarchical (cognitive radio)
Spectrum Sharing Models 21
-
Spectrum Property Right: Liberal License
¨ Spectrum is publicly owned, but licensed for exclusive use
¨ Liberal use rules ¤ Does not dictate technology
(cellular, WiFi commons, satellite,…)
¨ Allows spectrum trading ¤ Can re-allocate spectrum on large scales. ¤ Can define/trade spectrum contracts on finer scales
22
-
Spectrum Commons
¨ State-regulated (unlicensed) ¤ Spectrum owned by government ¤ Etiquette rules part of industry standard (802.11)
¨ Privately owned (licensed) ¤ Licensee sets rules, polices band ¤ Revenue from selling approved equipment
¨ Open access ¨ Requires etiquette rules
for sharing
23
-
Engineering Approach to Spectrum Crunch
¨ Add intelligence to mobile devices ¤ Frequency agility ¤ Wideband sensing ¤ Interference avoidance ¤ Adaptive quality of service
(context aware)
¨ Enables spectrum scavenging Cognitive Radio
Mitola and Maguire (1999)
24
-
Hierarchical
¨ Primary and secondary users ¨ Secondary users must not
disrupt primary users ¨ Relies on cognitive radio
25
-
Hierarchical
¨ State-regulated ¤ Spectrum owned by government ¤ Use rules for secondary users part of standard (802.22)
¨ Private contracts with “spectrum scavengers”
¤ Interference levels/payments set by mutual agreement
¨ Primary and secondary users ¨ Secondary users must not
disrupt primary users ¨ Relies on cognitive radio
26
-
Hierarchical: Technologies
¨ Underlay: low-power, spread spectrum for secondary users
¨ Overlay: exploit white spaces left by primary users
¨ Primary and secondary users ¨ Secondary users must not
disrupt primary users ¨ Relies on cognitive radio
27
-
Current Allocations
¨ Mix of: ¤ restricted use bands (e.g., broadcast TV) ¤ liberalized licenses (cellular) ¤ state-regulated commons (WiFi)
¨ Active trading of liberalized licenses among commercial service providers
¨ US policy trends have favored assignments of unlicensed spectrum over liberalized licenses
28
-
Why Unlicensed Spectrum?
¨ Pushed by DARPA
29
-
Why Unlicensed Spectrum?
¨ Pushed by DARPA ¤ Military needs distributed, dynamic methods for
spectrum sharing across military units
¨ Also by Microsoft, Google, Apple ¤ Facilitates 3rd-party software applications
¨ Success of WiFi
30
-
IEEE Spectrum Magazine, March 2004
Bold Predictions… 31
-
January 2014
Arguments for Unlicensed Spectrum
¨ Encourages innovation (WiFi, Bluetooth, etc.). ¨ Reduces barriers to entry. ¨ Has created at least as much value as licensed.
Huawei, Rolling Meadows
32
See [Benkler, 2012], [Eilat, Milgrom, Levin, 2011]
-
January 2014
Counter-arguments
¨ Technologies developed for unlicensed spectrum could be developed for licensed use as well.
¨ WiFi uses a simple, distributed interference management scheme inappropriate for applications requiring high spectral efficiency.
Huawei, Rolling Meadows
Unlicensed spectrum encourages innovation (WiFi, Bluetooth, etc.)
33
-
January 2014
Counter-arguments
¨ This may lead to a tragedy of the commons due to excessive interference.
¨ Unlicensed erects barriers to investment along with technologies and applications that may generate the highest social value.
Huawei, Rolling Meadows
Unlicensed spectrum lowers barriers to entry.
34
-
January 2014
Value of Unlicensed
Huawei, Rolling Meadows
Unlicensed spectrum has created at least as much value as licensed.
35
[According to FCC Commissioner Jessica Rosenworcel], “recent economic studies that add up the broader impact of unlicensed spectrum on the economy estimate its annual value at more than $140 billion.” FCC's Rosenworcel looks to 5 GHz band, 600 MHz guard bands for unlicensed wireless - FierceWireless
-
January 2014
Counter-arguments
¨ The value of spectrum must be separated from the value of the application it supports.
¨ The total value must be distinguished from the marginal value. ¨ Value depends on congestion and interference management.
Huawei, Rolling Meadows
Unlicensed spectrum has created at least as much value as licensed.
36
-
January 2014
How to assess spectrum value?
¨ Evaluate as an opportunity cost: If spectrum X were not available, then what would be the cost of serving those applications?
¨ Evaluate marginal value: If additional spectrum δX were allocated to application A, then how much additional demand could be supported?
¨ Evaluate congestion cost: How much is congestion reduced by adding spectrum or investing in interference management?
Huawei, Rolling Meadows
37
-
The Case for Liberal Licenses
ITA, San Diego February 2014
38
¨ Advocates: ¤ Encourages investment
(cellular), efficient use ¤ Does not constrain technology,
rules for spectrum sharing n Cellular, WiFi, Satellite,
broadcast… ¤ Facilitates trading
[Hazlett, 2010]
¨ Detractors: ¤ Has led to oligopoly
-
Spectrum Supply Curve
Spectrum price ($/Hz)
supply
demand
Qua
ntity
of s
pect
rum
(Hz)
As the spectrum price goes to zero: – The supply decreases due to the decrease in spectral efficiency. – The demand increases due to introduction of new services.
Equilibrium price
39
-
Commons vs Market
Spectrum price ($/Hz)
market transaction costs < cost of interference è Set up spectrum market
Qua
ntity
of s
pect
rum
(Hz)
Spectrum market Commons
supply
demand
p*
40
-
Commons vs Market
Spectrum price ($/Hz)
cost of interference < market transactions costs è Use commons model
Qua
ntity
of s
pect
rum
(Hz)
Spectrum market Commons
supply
demand
p*
41
-
Bit pipe Spot markets Local transactions
Market Structures 42
-
“Bit Pipe” Model
¨ “Wholesale” contract with cellular provider
Kindle
43
-
“Bit Pipe” Model
¨ “Wholesale” contract with cellular provider ¨ Mobile Virtual Network Operators (MVNO’s)
¤ Resells mobile services (e.g., Virgin Mobile)
44
-
“Bit Pipe” Model
¨ “Wholesale” contract with cellular provider ¨ Mobile Virtual Network Operators (MVNO’s)
¤ Resells mobile services (e.g., Virgin Mobile)
¨ Emerging model for wholesale cellular provider ¤ Open broadband network ¤ No retail services ¤ Wide coverage
45
-
“Bit Pipe” Model: Properties
¨ Wide-area coverage, high mobility ¨ Interference management ¨ Quality of Service guarantees ¨ Facilitates new wide-area wireless services ¨ Well-matched to lower frequency assignments
46
-
Spectrum Spot Market
Licensees A, B, C, …
Spectrum Broker
Service providers (Acme Wireless)
Service requests
¨ Immediate access, rapid (automated) transactions ¨ Low transaction costs ¨ Facilitates local services ¨ No need to build out large footprint
47
-
Wireless Spot Market
Owners A, B, C, …
Spectrum Broker
Service providers (Acme Wireless)
Service requests
¨ Spectrum can be bundled with equipment ¨ Broker allocates spectrum and technology
(e.g., cellular/commons) ¨ Spot market for bits
48
-
Market Mechanisms
¨ Sets prices, attempts to clear market
¨ Auction mechanism: collects bids; determines allocation
¨ Can be automated (“spectrum server”)
Owners A, B, C, …
Spectrum Broker
Service providers (Acme Wireless)
49
-
Spectrum Contracts
Owners A, B, C, …
Spectrum Broker
Service providers (Acme Wireless)
¨ Contracts can be arranged across: ¤ Frequency (spread spectrum, underlay) ¤ Locations (mesh networking) ¤ Time (time-of-day, futures, scavenging)
¨ Variable QoS guarantees (statistical)
50
-
Interference Management
Owners A, B, C, …
Spectrum Broker
Service providers (Acme Wireless)
¨ Spectrum assets predefined by owners ¤ Service providers must comply with power limits,
interference constraints ¨ Spectrum assets partitioned, assigned by broker
to satisfy service requests
51
-
Commercial broadband access Clean slate
How should spectrum be managed? 52
-
Spectrum Management: Observations
¨ Coverage is easy ¨ Interference
management is difficult
Low Freqs. (< 1 GHz)
High Freqs. (> 3 GHz)
¨ Coverage is difficult ¨ Interference
management is easy
Frequency à
53
-
Spectrum Management: High-Level
¨ Wide-area coverage ¨ Coordinated
interference management
¨ Expensive infrastructure
Low Freqs. (< 1 GHz)
High Freqs. (> 3 GHz)
¨ Local coverage ¨ Distributed
interference management (random access)
¨ Inexpensive
Frequency à
54
-
Shrinking Cells
Low Freqs. (< 1 GHz)
Frequency à
High Freqs. (> 3 GHz)
¨ Macro-cells ¨ Coordinated
interference management
¨ Expensive infrastructure
¨ Femto-cells/WiFi ¨ Distributed
interference management (random access)
¨ Inexpensive
55
-
Spectrum Sharing Models
¨ Exclusive use (licensed) ¤ Encourages investment
in infrastructure
¨ “Bit Pipe” (e.g., Light Squared)
Low Freqs. (< 1 GHz)
High Freqs. (> 3 GHz)
¨ Commons ¨ Public
(unlicensed)
¨ Private (licensed)
Frequency à
56
-
Spectrum Management: High-Level
¨ Macro-cells ¨ Coordinated
interference management
¨ Expensive infrastructure
Bit Pipe (< 1 GHz)
¨ Femto-cells/WiFi ¨ Distributed
interference management (random access)
¨ Inexpensive
Frequency à
Commons (> 3 GHz)
Dynamic Spectrum Markets?
¨ Micro-cells ¨ Light interference
management
¨ Spectrum servers
¨ Rapid transactions
57
-
What about Cognitive Radio?
¨ Enables spectrum scavenging ¤ Sense/exploit white spaces
¨ Could be applied to licensed or unlicensed spectrum ¤ Focus has been on unlicensed
access for secondary users
58
-
Where does cognitive radio fit?
Bit Pipe (< 1 GHz)
Frequency à
Commons (> 3 GHz)
¨ Coordinated interference management
¨ Expensive infrastructure
¨ Distributed interference management (random access)
¨ Inexpensive
59
-
Where does cognitive radio fit?
Bit Pipe (< 1 GHz)
Frequency à
Commons (> 3 GHz)
¨ Coordinated interference management
¨ Expensive infrastructure
¨ Distributed interference management (random access)
¨ Inexpensive
60
-
Where does cognitive radio fit?
Bit Pipe (< 1 GHz)
Frequency à
Commons (> 3 GHz)
¨ Coordinated interference management
¨ Expensive infrastructure
Need only simple dynamic channel assignment. Sensing, scavenging is probably unnecessary.
61
-
Where does cognitive radio fit?
Bit Pipe (< 1 GHz)
Frequency à
Commons (> 3 GHz)
¨ Coordinated interference management
¨ Expensive infrastructure
Dynamic Spectrum Markets
Scavenging spectrum across multiple service providers via a spot market may be useful, but…
62
-
Where does cognitive radio fit?
Bit Pipe (< 1 GHz)
Frequency à
Commons (> 3 GHz)
¨ Coordinated interference management
¨ Expensive infrastructure
Dynamic Spectrum Markets
Scavenging spectrum across multiple service providers via a spot market may be useful, but… ! SPs would control available spectrum ! “cognitive” means “frequency-agile”
63
-
Spectrum Management: Two Views
¨ Spectrum is abundant ¤ It is just poorly managed
¨ Spectrum will remain scarce ¤ Applications will be generated to use new spectrum ¤ Shift to cheaper, spectrally inefficient technologies
64
-
Transition to Spectrum Abundance
¨ Tradeoff between spectrum efficiency and power efficiency ¤ Shift emphasis towards low-power, inexpensive wideband
signaling techniques
¨ Efficiency becomes limited by transaction costs ¤ Spot markets with distributed interference management ¤ Transparent (standardized?) mechanism (pricing/auction) ¤ Wireless devices: frequency/technology agile, compatible
with spot market mechanism
65
-
Many Remaining Challenges
Policy Economics
Engineering
Interference management Incentives, efficiency Market design
66
-
Many Remaining Challenges
Policy Economics
Engineering
Interference management Incentives, efficiency Market design
Transition to spectrum markets??
67