Transcript Authers:dawn xiaodong song,david wagner xuqing tian publication

Timing Analysis of Keystrokes And Timing
Attacks on SSH
AUTHERS:DAWN XIAODONG SONG,DAVID WAGNER XUQING TIAN
PUBLICATION:USENIX SECURITY SIMPOSIUM 2001
PRESENTER: HIRAL CHHAYA FOR CAP 6133
OUTLINE……

Secure Shell (SSH) weaknesses

Analysis of user keystroke patterns

Attack using inter-keystroke timing

Performance evaluation

Countermeasures

Comments and conclusion
SSH- SECURED SHELL
Offers an encrypted channel and strong
authentication.
 Replaces telnet, rlogin.
 protects a network from attacks such as IP
spoofing, IP source routing, and DNS spoofing.
 If attacker tries to takeover the network SSh
diconnects.

SSH- SECURED SHELL
WEEKNESSES OF SSH

Two seemingly minor weaknesses:
 Padding:
Payload
1-8 bytes
Reveals approximate data size
 In
Packet
length
Padding
length
Random
Padding
Integrity
data
(MAC)
interactive mode, every individual keystrokes is
sent to remote machine in separate IP packet.
Leaks timing information of user’s typing
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
TRAFFIC SIGNATURE ATTACK
NESTED SSH ATTACK




User establish SSH
session between A and B
Another session from B
to C
User’s password for C is
transmitted across ssh
link A-> B
Adversary sniff the
password
STATISTICAL ANALYSIS OF INTER-KEYSTROKE
TIMINGS

Pick a pair of characters,
e.g. (“v”, “o”)

Ask users to type the pair
for 30-40 times

Collect latency information

Repeat for every different
pair of characters
HIDDEN MARKOV MODEL (HMM)
HMM - A Markov model when the current state
can not be observed.
 Outputs of the process are observed.
 Probability of output depends only on the state.
 Information on the prior path of the process
can be inferred from it’s output.

HIDDEN MARKOV MODEL (HMM)



Character pair is the hidden state.
Keystroke latency measured is the output observation.
Two assumptions:
 Probability of transformation from current state to another
state, depends only on the current state.
 Probability distribution of latency, depends only on the
current state
q = character pair
y = latency observation
BIG PICTURE
A
B
Detect SSH session
detect nested SSH
or SU
Sniffer
Keystroke Timing
statistics
n-Viterbi
Possibilities
Password
COUNTERMEASURES

Let the server return dummy packets when it
receives keystroke packets from the client.

Let the client randomly delay sending keystroke
packets.

Let the client send keystroke packets at a
constant rate.
STRENGTHS

Novel idea

Nice technique

Good performance

Interesting findings

Countermeasures given
WEEKNESSES

No mention of how to deal with backspace

No discussion of how different keyboard layouts
affect the results
 Laptop
vs desktop
 Different
keyboard layouts in different regions
THANK YOU