Opportunistic Spectrum Access:
Challenges, Architecture,
Protocols
C. Santivanez, R. Ramanathan (presenter), C. Partridge, R.
Krishnan, M. Condell, S. Polit
Internetwork Research Department
BBN Technologies, Cambridge, MA
(ramanath@bbn.com)
Outline
• Introduction and motivation
• Spectrum Agility
• Policy Agility
• A simple system for opportunistic spectrum access
2
Opportunistic Spectrum Access
• Basic idea
– Sense the spectrum you want to transmit in
– Look for “holes” or “opportunities” in time and frequency
– Transmit so that you don’t interfere with the licensees
WALL
PRIMARY 1
PRIMARY 2
FREQUENCY
PRIMARY 3
PRIMARY 4
SECONDARY
TIME
HOLE
3
Problem Statement
• Given
Frequency
– a set of possibly mobile nodes
comprising a wireless network
Time
– the hole/wall characteristics as a
Frequency
function of time for each node
– a set of constraints on the spatial
Time
reuse of frequencies
Frequency
– [a set of capacity demands per link]
Time
• dynamically control access
(xmt/rcv, beam, power etc.) on
Frequency
each frequency such that
– network capacity is maximized Time
– constraints are honored
– [capacity demands are met]
Frequency
Time
4
Why Opportunistic Spectrum Access
• Current “static” spectrum allocation
strategy is wasteful
– Huge opportunities exist in time,
frequency and space
– Apparent spectrum scarcity
• Deployment difficulty
– Allocating spectrum for overseas
military operations
– Using WiFi in different countries in Maximum Amplitudes
conformance with the country’s policy Heavy Use Heavy Use
Less than 6% Occupancy
Amplidue (dBm)
• Spectrum policy is simply outdated Sparse Use Medium Use
– FCC Spectrum Policy Task Force
Frequency (MHz)
5
OSA in the Wireless Internet: Drivers
• Need for more capacity
– 2G, 3G, 4G……
• Waveform Diversity
– Beamforming, MIMO, UWB, OFDM, Spectrum-adaptive/heteromorphic …..
• Software/Agile/Cognitive Radios
– JTRS, Vanu Inc., GNU Radio ……
• Secondary markets, FCC Spectrum Policy, ….
Current trends point to a need for and capability to do opportunistic
access as a key part of next generation Wireless Internet architecture
6
OSA Challenges
• Spectrum Agility
– How do we identify holes (opportunities) ?
– How can we access those opportunities?
– How do we prevent interfering with primary users?
• Policy Agility
– How do we control access to be in conformance with regulatory policies?
– How can we effect a change in node behavior in accordance with policy changes?
– How do we support multiple (concurrent) policy authorities
7
Next …..
• Introduction and motivation
• Spectrum Agility
• Policy Agility
• A simple system for opportunistic spectrum access
8
Spectrum Agility Components
Opportunity Awareness
Determine spectrum opportunities
and usage constraints, based on
sensing, and policy information
Opportunity
Info, and Usage
constraints constraints
Opportunity Allocation Opportunity Use
Determine how the available Per-packet Transmit packet in accordance with
opportunities are shared on an assignment opportunity assignment, adjust
instant-by-instant basis parameters as per constraints
9
Challenge: Opportunity Awareness
• Wideband sensing
– Problem: Sense energy/ receive signatures over a large portion of the spectrum
(e.g. entire U-NII indoor band (5.15 - 5.25 GHz (100 MHz))
– Current wireless receiving technology receives/transmits over much smaller
chunks (e.g. each channel in 802.11b is about 22 MHz)
– An alternative: Primaries log their current/planned use, secondaries access this
database
• Opportunity Identification
– Problem: Based on sensing, decide whether or not to use a slice of spectrum
– Interference occurs at receivers, but you can only sense transmitters
– Hidden transmitters
• Opportunity Dissemination
– Problem: Share opportunities with other nodes to coordinate use of spectrum
– Opportunity dissemination can be bandwidth intensive
– Chicken-and-egg: Which channel do we use to disseminate opportunities?
10
Challenges: Opportunity Allocation/Use
• Allocation 4,5,6,7
1,2,3,4,5
– Problem: Channel access over 5,6
dynamically changing holes 1,2
X RTS (3,4,5) Y
– CSMA/CA: Need to acquire floor over CTS (3,4)
1,2,3,4,5 3,4,5,6
a common channel set Z
– TDMA: Assign time slots over each 1,2,3,4
available frequency
• Use
– Problem: Implementing the allocation 5
B-C C-D C-D C-D
– Waveforms that can use B
1,2,3,4,5 A
1,2,3
4 B-C C-D B-C B-C
discontiguous portions of the spectrum 4
4 3 5 3 B-E
– Dynamically adapt waveform to
A-C A-C A-C
A-D
2 B-E A-D A-D
required PSD 4
A-D A-D B-E B-E
E C D 1
– Adjust beamforming, data rate, 2,3,4 3,4,5
2,3,4,
modulation, spreading … 5 S1 S2 S3 S4
11
Next …..
• Introduction and motivation
• Spectrum Agility
• Policy Agility
• A simple system for opportunistic spectrum access
12
Policy Agility
FCC Rule Book Limited or no
Hardwired policy programmability
(e.g. ASICs)
Canned behaviors:
few/fixed modes
of operation
Machine-Readable Policies
Highly Programmable,
Dynamic Policy fast, low power
devices (e.g FPGAs)
Agile behaviors:
numerous modes
of operation
Policy agility is necessary to exploit the emerging agility of devices
and allow in-situ policy-based control of radio behaviors
13
Machine Understandable Policies
OSA node
Machine-readable
Policies
Spectrum OSA node
Policies
Internet
Policy
Repository
14
Benefits of policy agility
• Adaptation to policies changing over time
– Allows development of technology in advance of policies
• Adaptation to policies changing over geography
– E.g., use a new smart card when in a new country
• Self-checking policies
– Implications of policy interactions can be worked out in advance
– Consistency checking at the logical level
• Sub-policy management
– Secondary markets: Overarching policy, sub-policy – allow primary user to develop
sub-policies for secondary users
• Radio system capabilities based policies
– E.g. two different policies depending on whether a radio system can detect certain
waveforms or not
15
Challenges: Policy Agility
• A language for policy expression
– Inheritance, Reification (rules about rules), Inference (derivable rules),
Extensibility, Scalability, Declarative
• An ontology for policies
– What are the “primitive” objects underlying policies?
– What are the inter-relationships between the primitive objects
• Reasoning about policies
– Conformance: Is this usage permitted by policy?
– Constraints: What is the range of parameters for this particular usage?
– Searches: Here are my specs, my needs. Give me the usage.
• BBN has developed an initial ontology and a policy language based
on OWL (Semantic Web Language)
16
Next …..
• Introduction and motivation
• Spectrum Agility
• Policy Agility
• A simple system for opportunistic spectrum access
17
A policy-aware OSA system
Rest of talk
Ask for transmission opportunities
System
System
Policy Strategy Return transmission opportunities
Device
Configuration
Reasoner
OSA
capable
Device
Sense
Policy
Regulatory Conformance Ask for usage validity
Policy
Reasoner Transmit
Allow/Deny
18
Spectrum Agility: Components
To SSR To PCR
Information Flow
Control Flow Upper Layer
XG Opportunity API
Sense Interface Kernel ND-HIP
Nbr list, nbr HIAs
OSA My HIA - Waveform
Adaptation Opportunity selection - ND
My IDLE CHANNEL,
Layer Identification - Idle Channel my HIA, pathloss - HIP
Selection - RMAC
FFT
XG Transceiver API
TRANSCEIVER
19
Sense Interface
Hole Information Array (HIA)
1 1 0 0 0 1 0 0 1 1
Frequency (KHz)
20
IDLE Channel Selection
– IDLE Channel (freq., rate, waveform) is not the XG
Coordination Channel
– While node listen to IDLE Channel MUST be reachable by all
1-hop neighbors.
Idle Channel
My_HIA
1 1 0 0 0 1 0 0 1 1
Nbr_1_HIA
1 0 0 0 0 1 1 0 0 1
Nbr_2_HIA
0 1 1 0 0 0 1 0 1 1
Nbr_3_HIA
1 1 0 0 0 1 0 1 1 1
21
IDLE Channel Selection (contd.)
– Communication with neighbors may then switch to a more
favorable channel.
– E.g. RTSs need to be sent using a nbr IDLE channel, while
DATA packet may be sent using one of the channels below:
Point to point opportunity to Nbr 1
My_HIA
1 1 0 0 0 1 0 0 1 1
Nbr_1_HIA
1 0 0 0 0 1 1 0 0 1
My_HIA
1 1 0 0 0 1 0 0 1 1
Nbr_2_HIA
0 1 1 0 0 0 1 0 1 1
Point to Point opportunity to Nbr 2
22
BW Occupation (x 100 KHz) BW Occupation (x 100 KHz)
Time
Time
Zoom
Spectrum Utilization
23
Concluding Remarks
• Opportunistic Spectrum Access de-bunks the “spectrum scarcity”
myth and a promising technology for Wireless Internet
• Realizing true opportunistic spectrum access requires solving not only
the spectrum agility problem but also the policy agility problem
• There is plenty of “low hanging fruit” that can be harvested with a
relatively simple set of mechanisms
– We presented one such system, that incorporates both spectrum and policy agility
– Showed an order of magnitude gain in capacity
• Future research to target “higher hanging fruit”
– Dealing with rapidly changing hole information in a large network
– Constructing a coordination channel “on the fly”
– Theoretical capacity bounds of OSA
24
Resources
• Much of this work was based on the DARPA XG Program
– Phase I: BBN project “Medium Access Control for XG (X-MAC)”
– Phase 2: BBN project “XG Architecture and Protocols (XAP)”
• Documents
– “The XG Vision”, version 2.0, http://www.ir.bbn.com/projects/xmac/vision.html
– “The XG Policy Language Framework”, version 1.0, http://www.ir.bbn.com/projects/xmac/vision.html
• Web sites
– http://www.ir.bbn.com/projects/xmac/index.html
– http://www.darpa.mil/ATO/programs/xg/index.htm
25