Practical Known-Plaintext Attacks against Physical Layer Security in - - PowerPoint PPT Presentation

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Matthias Schulz, Adrian Loch, Matthias Hollick

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

2

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Motivation

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Application Transport Network Data Link Physical

Cryptography

computational security

Physical Layer Security

aims at information-theoretical security no computational restrictions on eavesdropper powerful attack models

3

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Motivation

STROBE: Orthogonal Blinding

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

STROBE: Actively Securing Wireless Communications using Zero-Forcing Beamforming

Narendra Anand

Rice University Houston, USA Email: nanand@rice.edu

Sung-Ju Lee

Hewlett-Packard Laboratories Palo Alto, USA Email: sjlee@hp.com

Edward W. Knightly

Rice University Houston, USA Email: knightly@rice.edu

Abstract—We present the design and experimental evalua- tion of Simultaneous TRansmission with Orthogonally Blinded Eavesdroppers (STROBE). STROBE is a cross-layer approach that exploits the multi-stream capabilities of existing technologies such as 802.11n and the upcoming 802.11ac standard where multi- antenna APs can construct simultaneous data streams using Zero- Forcing Beamforming (ZFBF). Instead of using this technique for simultaneous data stream generation, STROBE utilizes ZFBF by allowing an AP to use one stream to communicate with an intended user and the remaining streams to orthogonally “blind” (actively interfere with) any potential eavesdropper thereby preventing eavesdroppers from decoding nearby transmissions. extensive experimental evaluation, we sistently outperforms Omnidir (SUBF), and by

upcoming 802.11ac1 employ physical layers (PHYs) that can implement ZFBF to construct multiple parallel transmission streams to a single user (11n) or simultaneously to multiple users (11ac). Because such existing technologies are already able to create multiple parallel streams, STROBE can be implemented in these systems with minor AP no client modification. STROBE encryption methods

§ Published at INFOCOM 2012 § Practical Orthogonal Blinding implementation § Eavesdropper limited to one antenna

4

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Motivation § Introduction to Orthogonal Blinding § Contribution: Known-Plaintext Attack § Evaluation § Conclusion

Contents

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

5

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Motivation § Introduction to Orthogonal Blinding § Contribution: Known-Plaintext Attack § Evaluation § Conclusion

Contents

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

6

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

From Shannon to Wyner

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice Bob Eve channel channel encoder decoder M M Xn Yn Zn Degraded Wiretap Channel according to Wyner à Secrecy measured as information leakage to Eve

7

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

How to reduce information leakage?

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice Bob Eve channel channel encoder decoder M M Xn Yn Zn Degraded Wiretap Channel according to Wyner The channel to Eve should introduce additional noise channel

8

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Orthogonal Blinding

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice Bob Eve channel encoder decoder M M Xn Yn Zn The channel to Eve should introduce additional noise AN f(M,AN) Artificial Noise (AN) transmitted orthogonally to Bob’s channel: “blinding” only Eve channel channel

9

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Orthogonal Blinding

Practical Implementation

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice

multi-antenna node

Multiple Eves

multiple single-antenna nodes

Bob

single-antenna node Data Artificial Noise Filter Data Noise Noise Noise

10

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Motivation § Introduction to Orthogonal Blinding § Contribution: Known-Plaintext Attack § Evaluation § Conclusion

Contents

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

11

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Known Plaintext Attack

System Model

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice

multi-antenna node

Eve

multi-antenna node OR multiple cooperating single-antenna nodes

Bob

single-antenna node Data Data Noise Data Known by Eve Adaptive Filter Data Artificial Noise Filter

12

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Known Plaintext Attack

System Model

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice

multi-antenna node

Eve

multi-antenna node OR multiple cooperating single-antenna nodes

Bob

single-antenna node Data Data Noise Data Known by Eve Adaptive Filter Data Artificial Noise Filter Adaptive Filter filter

• utput

ω0 ω1

Known Data Filter Update Calculation

(LMS or NLMS with step-size µ)

• ant. 1
• ant. 2

Evaluation

13

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Known Plaintext Attack

Noise to Data Ratio

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice

multi-antenna node

Eve

multi-antenna node OR multiple cooperating single-antenna nodes

Bob

single-antenna node Data Noise Data Known by Eve Adaptive Filter Data Filter low NDR med. NDR high NDR Noise to Data Ratio (NDR) Evaluation

14

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Known Plaintext Attack

Noise introduced by Wireless Channel

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Alice

multi-antenna node

Eve

multi-antenna node OR multiple cooperating single-antenna nodes

Bob

single-antenna node Data Noise Data Known by Eve Adaptive Filter Data Filter low NDR med. NDR high NDR Noise to Data Ratio (NDR)

Noise introduced by the wireless channel Signal to Noise Ratio (SNR)

Evaluation

15

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Motivation § Introduction to Orthogonal Blinding § Contribution: Known-Plaintext Attack § Evaluation § Conclusion

Contents

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

16

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Evaluation

Testbed

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

WARPLab and MATLAB Alice Bob Eve

17

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Evaluation

Eve’s Filter Convergence (measurement)

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

20 40 60 80 100 120 140 10−0.5 100 Number of training symbols Eve’s Symbol Error Rate µNLMS = 0.1 µNLMS = 0.3 µNLMS = 0.9

Noise to Data Ratio (NDR) = 4 Symbol Error Rate converges for certain number of training symbols minimum achievable Symbol Error Rate after convergence step-size of the Normalized Least Mean Squares adaptive filtering algorithm

18

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

Evaluation

Convergence performance (measurement)

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

10−1.4 10−1.2 10−1 10−0.8 10−0.6 10−0.4 10−0.2 20 40 60 80 2 4 6 8 10 2 4 6 8 10 Eve’s Symbol Error Rate at convergence Convergence time in training samples µNLMS = 0.3 µNLMS = 0.9 µNLMS = 1.4 Bob

higher step-size à faster convergence Noise to Data Ratio (NDR) lower step-size à lower Symbol Error Rate For comparison: Bob’s Symbol Error Rate

19

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

2 4 6 8 10 12 14 16 18 20 22 24 10−1 100 Number of Eve’s antennas Eve’s Symbol Error Rate SNRTX = 10 dB SNRTX = 25 dB Evaluation

Many eavesdropper antennas (simulation)

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

100 training symbols, Noise to Data Ratio (NDR) = 10, filter step-size: µNLMS = 0.3 more antennas à lower SER filtering complexity increases linearly

find additional results in our paper

20

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Motivation § Introduction to Orthogonal Blinding § Contribution: Known-Plaintext Attack § Evaluation § Conclusion

Contents

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

21

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014

§ Successful secrecy reduction § Adaptive filtering used for known-plaintext attacks § Simulation and experimental evaluation

Conclusion

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

If you ever propose a physical layer security scheme à à à consider multi-antenna eavesdroppers ß ß ß

22

Matthias Schulz, Adrian Loch, Matthias Hollick – NDSS 2014 Matthias Schulz

Department of Computer Science SEEMOO

• Mornewegstr. 32

64293 Darmstadt/Germany mschulz@seemoo.tu-darmstadt.de Phone +49 6151 16-70928 Fax +49 6151 16-70921 www.seemoo.tu-darmstadt.de

Practical Known-Plaintext Attacks against Physical Layer Security in Wireless MIMO Systems

Thank you for your attention