performance modeling of anonymity protocols

40
Performance Modeling of Anonymity Protocols Carey Williamson Niklas Carlsson Andreas Hirt Michael J. Jacobson, Jr. Department of Computer Science University of Calgary Financial support for this research support was provided by: Natural Sciences and Engineering Research Council (NSERC), Informatics Circle of Research Excellence (iCORE), Alberta Ingenuity Fund (AIF), and Canada Foundation for Innovation (CFI)

Upload: elu

Post on 08-Jan-2016

23 views

Category:

Documents


0 download

DESCRIPTION

Performance Modeling of Anonymity Protocols. Carey Williamson Niklas Carlsson Andreas Hirt Michael J. Jacobson, Jr. Department of Computer Science University of Calgary Financial support for this research support was provided by: - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Performance Modeling of Anonymity Protocols

Performance Modelingof Anonymity Protocols

Carey Williamson Niklas Carlsson

Andreas Hirt Michael J. Jacobson, Jr.

Department of Computer Science

University of Calgary

Financial support for this research support was provided by:

Natural Sciences and Engineering Research Council (NSERC),

Informatics Circle of Research Excellence (iCORE),

Alberta Ingenuity Fund (AIF), and

Canada Foundation for Innovation (CFI)

Page 2: Performance Modeling of Anonymity Protocols

2

Introduction

Anonymous communication conceals who communicates what, to whom, and when

Allows individuals to communicate without fear of embarrassment, ridicule, or retribution

Cornerstone for freedom of speech

Page 3: Performance Modeling of Anonymity Protocols

3

Some Real World Applications

Good: Freedom of speech in totalitarian regime Crime stoppers On-line counseling Whistle blowing Group evaluations Military communications …

Bad: Organized crime Terrorist groups ...

Page 4: Performance Modeling of Anonymity Protocols

4

Outline

Review of Anonymity Schemes Our Work: Buses, Taxis, Motorcyles Performance Modeling Numerical Results Conclusion

Page 5: Performance Modeling of Anonymity Protocols

5

xmkzykrziwqmqkdx

Re-routing with Layered Encryption

Layered Encryption: Add layers of encryption to make message contents change each hop

hello

Page 6: Performance Modeling of Anonymity Protocols

6

xmkzykrziwqm

Layered Encryption: Add layers of encryption to make message contents change each hop

hello

Re-routing with Layered Encryption

Page 7: Performance Modeling of Anonymity Protocols

7

Layered Encryption: Add layers of encryption to make message contents change each hop

hello

xmkzykrz

Re-routing with Layered Encryption

Page 8: Performance Modeling of Anonymity Protocols

8

Layered Encryption: Add layers of encryption to make message contents change each hop

hello

xmkz

Sender?

hello

Problem: Timing analysis

Re-routing with Layered Encryption

Page 9: Performance Modeling of Anonymity Protocols

9

Senders use nested (layered) encryption along re-routing path Mixes (re-routing nodes) mix input-output correlations:

Collect input batch Peel encryption layer away Output in random order

Message 1

Message 2

Message 3

Message 4

Message 5

Message 4

Message 3

Message 1

Message 5

Message 2

Mixes

Page 10: Performance Modeling of Anonymity Protocols

10

Current Solutions

No Cover Traffic

Partial Cover Traffic

Full Cover Traffic

Schemes Crowds, TOR JAP, MorphMix

Mixmaster, Mixminion, Tarzan

Anonymity Weak Moderate Strong

Problems Vulnerable to known attacks

Vulnerable to known attacks

Not suitable for interactive applications, don’t scale well

Page 11: Performance Modeling of Anonymity Protocols

11

Classic Buses Protocol[Beimel and Dolev 2003]

Metaphor: city bus, with regularly scheduled route, which obscures the movements of its messengers

Assume dark windows, and enclosed garages at each stop

hello

hello

Page 12: Performance Modeling of Anonymity Protocols

12

Anonymity in Buses

Sender Anonymity: Suspected sender can claim they are forwarding a message on behalf of any other participant on the bus path

Receiver Anonymity: Suspected receiver can claim they forwarded a message to any other participant on the bus path

Page 13: Performance Modeling of Anonymity Protocols

13

Key Ideas in Our Buses

Indirection path: re-routing path on top of bus overlay

Layered Encryption: encryption on reverse indirection path

Owned Seats: Each participant replaces owned seats every bus tour (online)

Receiving seats: bus copied and decrypted offline to find messages

Page 14: Performance Modeling of Anonymity Protocols

14

Buses Protocol

S Rhello

Page 15: Performance Modeling of Anonymity Protocols

15

S Rhello

xmkz

Buses Protocol

Page 16: Performance Modeling of Anonymity Protocols

16

S Rhello

ymkq

Buses Protocol

Page 17: Performance Modeling of Anonymity Protocols

17

S Rhello

Buses Protocol

Page 18: Performance Modeling of Anonymity Protocols

18

S Rhello

Buses Protocol

Page 19: Performance Modeling of Anonymity Protocols

19

ymkq

S Rhello

xmkz

Buses Protocol

Page 20: Performance Modeling of Anonymity Protocols

20

helloxmkz

S Rhello

Buses Protocol

Page 21: Performance Modeling of Anonymity Protocols

21

Improvements with Taxis

Processing Delay decreased by O(n) Owned seats are delayed once per bus tour

instead of n times (see MASCOTS 2008 paper )

Networking Delay decreased by O(n) Forwarding of unowned taxis can be

pipelined by giving unowned taxis network priority over owned taxis (see MASCOTS 2008 paper)

Page 22: Performance Modeling of Anonymity Protocols

22

Improvements with Motorcycles

Routing Path length decreased to O(log n) Chord-based routing using finger table

Forwarding delay actually increases More “message transfers” occur at nodes

Still a net win overall!

Page 23: Performance Modeling of Anonymity Protocols

23

Performance metric: one-way message delay DSR

Five main components Sender S must create/encrypt and send message Load-dependent sender-side delay

Queueing of (average) duration Ws Load-independent path delay

Path length HSR with (Dproc+Dnet) delay on each node Load-dependent transfer delay

Queueing at HT transfer nodes, each with duration WT

Target receiver R must decrypt and receive message

Model Overview

RTTnetprocSRSSSR DWHDDHWDD )(

Page 24: Performance Modeling of Anonymity Protocols

24

Load-independent Delays

AnonymityProtocol

ProcessingDproc

NetworkDnet

Buses KNDseat KNs/r+p

Taxis KDseat Ks/r+p

Motorcycles

KDseat Ks/r+p

N nodes; K seats per node; Dseat processing per seat; s/r transmission time per seat; p per-hop propagation delay

Page 25: Performance Modeling of Anonymity Protocols

25

Hop counts

Metric Buses/Taxis Motorcycles

HSR

(end-to-end)

N/2, if L=0

(1+L)(N+1)/2, otherwise

HT

(transfers)L HSR – 1

2

log)1(

log

1

1 log1

NL

h

Nh

N

L Nh

Page 26: Performance Modeling of Anonymity Protocols

26

Load-dependent Delays

ProtocolSender

WS

TransfersWT

Cycle TimeTC

Buses

Taxis

Motor )(2

log)2/1()2/1(

)()2/1()2/1(

)()1()2/1(

pr

KsKD

NQTQT

pr

KsKDNQTQT

pr

KNsKNDNQTQT

seatCCCC

seatCCCC

seatCCCC

Page 27: Performance Modeling of Anonymity Protocols

27

Light Load Case

Light load: No queueing QC 0 Example: Buses protocol

Dproc ~ N; Dnet ~ N; TC ~ N2; hence, DSR ~ N2

Scaling behavior Buses: DSR ~ N2

Taxis: DSR ~ N Motorcycles: DSR ~ log2N

RCnetprocSSR DTLDDN

LDD )2

1()(

2)1(

Page 28: Performance Modeling of Anonymity Protocols

28

Queueing Analysis (1 of 3)

Single-seat (K=1) case Analysis on per-node basis

New messages at rate /N Message transfers at rate HT/N Assume Poisson arrivals at aggregate rate (1+ HT)/N

Either: - service period of duration TC

- vacation period of duration TC

(1+ HT)/N

Node i

Page 29: Performance Modeling of Anonymity Protocols

29

Can be shown that generating function

In our system

Queueing Analysis (2 of 3)

1

001

,)(

))(1)(()(

zdz

dFpwhere

zzV

zFzVpzQ

NH

ezvzvzFzV

T

z

i

ii

i

ii

λ)1(

...)()( )1(

00

Page 30: Performance Modeling of Anonymity Protocols

30

Queueing Analysis (3 of 3)

Expected queue length

Other metrics “relatively straightforward” to obtain, given the generating function Variance State probabilities q0, q1, …, qm

)1(2

)23(...][

10

zii dz

dQiqqE

2q

Page 31: Performance Modeling of Anonymity Protocols

31

Experimental Validation (Buses)

Page 32: Performance Modeling of Anonymity Protocols

32

Experimental Validation (Taxis)

Page 33: Performance Modeling of Anonymity Protocols

33

Simulation Validation (Buses)

Page 34: Performance Modeling of Anonymity Protocols

34

Simulation Validation (Taxis)

Page 35: Performance Modeling of Anonymity Protocols

35

Simulation Validation (Motorcycles)

Page 36: Performance Modeling of Anonymity Protocols

36

N=4 N=16

Impact of message generation rate

Different saturation points ( 1) E.g., capacity planning

Page 37: Performance Modeling of Anonymity Protocols

37

Buses

Taxis

Motorcycles

Impact of node utilization

Queueing delays dominate when > 0.8

Note higher saturation point …

can sustain higher Hence, differences even

greater than shown …

Page 38: Performance Modeling of Anonymity Protocols

38

Buses

Taxis

Motorcycles

Scaling results

for light load with K seats per node

Low load results As expected, scales

as (roughly) Buses N2

Taxis N Motorcycles log2N

Page 39: Performance Modeling of Anonymity Protocols

39

Buses

Taxis

Motorcycles

Scaling results

for different load levels

Relative performance differences maintained at higher loads

In summary: Motorcycles provide a robust and scalable approach for anonymous network communication.

Page 40: Performance Modeling of Anonymity Protocols

40

Conclusions

The average message latency of Practical Buses scales quadratically with number of participants

Analysis, simulation, and experimental results The average message latency of Taxis scales

linearly with the number of participants Analysis, simulation, and experimental results

The average message latency of Motorcycles scales logarithmically with the number of participants

Analysis and simulation results