Professional Documents
Culture Documents
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Material
Interactions of CRs Impact of Hostile Users Security Issues
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Housing Bubble
Bounce up instead of down Slower interactions lead to slower changes Also indicative of the role beliefs play in instability
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
In heavily loaded networks, a single vacation p p process can spawn an infinite adaptation p
Suppose
g31>g21; g12>g32 ; g23>g13
2
Infinite Loop!
4,5,1,3,2,6,4,
1
Interference Characterization
Chan. Interf. ( , , ) (0,0,0) ( , , ) (0,0,1) ( , , ) (0,1,0) ( , , ) (0,1,1)
( , , ) (1,0,0)
( , , ) (1,0,1)
( , , ) (1,1,0)
( , , ) (1,1,1)
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Our use:
Every time a cognitive radio unilaterally adapts in a way that furthers its own goal, goal some real-valued function increases. Predictable, stable emergent behavior Behavior inconsistent with the goals will immediately break the monotonicity
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
V(a)
time
Example behavior
For example, for a collection of 802.11 clusters independently choosing operating frequencies 1. All self-interested adaptations
Scalable resource utilization No synchronization required
0 5
Clusters Frequencies
10 15 20
25-30 dB
25-30 dB
No network is an island
Many TVWS standards
802.22 802 22 (CR for rural ) 802.16h (CR WiMAX) 802.11af (WhiteFi) CogNeA Decisions impact one another
Etiquette
Share information
Sensing
Merging? M i ?
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Suppose another network is compromised in your area Their behavior influences your networks adaptations
Stability impact
Clusters Frequencies
0 5 10 15 20
~30 dB
pe sloy) ter ilit reamob G (
CognitiveCRT Proprietary Radio Technologies 147 Mill Ridge Rd, Ste 119 9 Lynchburg, VA 24502
Hostile CR
35 normal
Adapt at inopportune times Simply minimize performance Ensure marginally stable E i ll t bl network goes unstable Plus learning exploits
And spoofing p g
And information corruption
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Malicious != Selfish
Popular solution to mischievous nodes (selfish nodes that ( f damage network) is to punish nodes
Also implies a way to brainwash learning nodes
From Fig 6 in [MacKenzie_01]
Imperfect information can obfuscate punishment bf t i h t from mischievous behavior and produce catast op c catastrophic cascades
Brittleness
Adaptations
Radio Utilities
Radio Utilities
-50
f) (f (dBm)
(f) (dBm)
-55
Potential Function
50 100 150 iteration 200 250 300 Web: www.crtwireless.com
Cognitive Radio Technologies 100 0 50 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Implementation Discussion
Implement as monitoring system that evaluates potential (emergent) f nction function
Frequently sum of performance levels Complexity is in the transmission / connectivity t i i ti it No single node / cluster knows / can evaluate emergent function
With BSI, a malicious node t S, a c ous ode cant tell a credible lie!
Other relationships exist Need to be WPG / EPG for linear relationships
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
Security Issues
[Clancy_08] Primary user emulation attacks B li f manipulation attacks Belief i l ti tt k A cognitive radio virus Spectrum sensing data falsification [Chen_08a] Quiet period jamming [Bian 08] [Bian_08] Replay sensing attacks [Bian_08] False coexistence information [Bian_08] [Bian 08] Honeypot attacks [Newman_09] Chaff point attacks [Newman_09]
[Brown_08]
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502
If there are vulnerabilities, how will you detect that they are being exploited?
Cognitive Radio Technologies 147 Mill Ridge Rd, Ste 119 Lynchburg, VA 24502 Web: www.crtwireless.com Ph: (540) 230-6012 Email: info@crtwireless.com