Você está na página 1de 110

Wireless Network Security

LECTURE 1: Introduction.
Wireless communication.
Security and privacy goals.
Brief overview of security
mechanisms.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

ACKNOWLEDGEMENTS

Slides used in these lectures are based on


Lecture slides I used at Purdue (555, 590T, 590W)
www.cerias.purdue.edu/homes/crisn

Slides provided with the book "Security and


Cooperation in Wireless Networks", J-P. Hubaux
and Levente Buttyan, 2008
http://secowinet.epfl.ch/index.php?page=slideshow.html

Research presentations from collaborations with


former students and colleagues
PhD Theses of Jing Dong and Reza Curtmola
NIST SP documents
wikipedia
Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Topics

Monday: Wireless communication


characteristics; Wireless networks and
applications. Security goals and adversarial
models
Tuesday: Attacks on protocols. Cellular
networks; WiMAX networks
Wednesday: 802.11 networks; Data link
security; Secure routing in ad hoc networks
Thursday: Sensor networks, Bluetooth
Friday: RFID, VANETs

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Readings

Security and cooperation in wireless


networks. L Buttyan, J-P Hubaux.
Slides
Research papers
NIST SP

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Today

Wireless
communication
Wireless architectures
Introduction to crypto
and attacks

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Security Services (or Goals)

1) Confidentiality: information is available for


reading only to authorized parties.
Example: Alice sends a message to Bob, only Alice and
Bob can understand the content of the message.

2) Authentication:
Data source authentication: the data is coming from
an authorized party.
Example: Alice receives a message from Bob. This
service ensures that the message is from Bob and not
from Carl.
Entity authentication: the entity is who it says it is.
Example: When Alice tries to obtain access to her
bank account, an authentication operation is performed
to ensure that Alice asks for the information.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Security Services (2)

3) Integrity: detect if data was modified, from the


source to the destination.
Example: Alice sends an email to Bob. Carl
intercepts the message and modifies it. Data
integrity allows for Bob to detect that the message
was modified on the way from Alice to him.
4) Non-repudiation: neither the sender, nor the
receiver of a message are able to deny the
transmission.
Example: Alice sends Bob a contract, signed. The
non-repudiation service ensures that Alice can not
claim that the signature was produced by
somebody else.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Security Services (3)

5) Access control: only authorized


parties can use specific resources.
Example: Alice wants to print a document, she must
be authorized to get that document and to use the
printer.

6) Availability: resources available to


authorized parties.
Example: A web site might become unavailable if
the server crashes, or is bombarded with requests.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Privacy

1) Anonymity: hiding who performed a given


action (k-anonymity)
2) Untraceability: making difficult for an
adversary to identify that a given set of actions
were performed by the same subject
3) Unlinkability: hiding information about the
relationships between any item
4) Unobservability: hiding of the items
themselves
5) Pseudonymity: making use of a
pseudonym instead of the real identity

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

Security and Wireless Communication

The usual goals: authentication,


confidentiality, integrity, access control,
non-repudiation, denial of service
Additionally: privacy is a big concern
particularly location privacy
Is wireless communication more
vulnerable?

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

10

Wireless Specific

Physical security: an issue for small devices,


Protocols have to consider inside attacks (assume the
device is controlled by the adversary)

Eavesdropping: easier than wired

Resource constraints: medium is shared, some


devices have limited power and computation
resources
Solutions must not have significant overhead
Privacy: many devices equipped with location
services, location privacy a bigger concern
Denial of service: jamming is an issue for some type
of wireless communication

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

11

Wireless communication
characteristics

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

12

Wireless Communication
THERE IS NO LINK: electromagnetic waves

c
f =

Cristina Nita-Rotaru

c (speed of light) = 3x108 m/s


f is the frequency
is the wavelength

Lecture 1/ WS Milano Summer 2011

13

Wave Propagation

Reflection: wave gets reflected when


it hits an object very large compared
with the wavelength
Diffraction: wave bends at the edges
and propagates in different directions
when it hits an impenetrable object
Scattering: wave scatters when it
travels through a medium containing
objects smaller than the wavelength
(e.g. trees)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

14

Characteristics

Interaction with the environment: path


loss, interference, blockage
Transmission constraints: constrain the
ability to transmit information at different data
rates
Noise: Quality of communication is affected
by noise
Error rate: Higher than in wired
communication

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

15

Wave Components

Direct path component:


Path loss: attenuation of an electromagnetic wave in
transit from a transmitter to a receiver
Fading: time variation of the received signal power
caused by changes in the transmission medium or
path
Doppler shift: the apparent change in frequency and
wavelength of a wave that is perceived by an observer
moving relative to the source of the waves

Multi-path components:
Result from reflection, refraction, scattering
Wave arrives at receiver shifted in amplitude, phase
and frequency

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

16

Path Loss

Power of transmission reduces:


terrain contours
different environments (urban or rural, vegetation
and foliage)
propagation medium (dry or moist air)
distance between the transmitter and the receiver
height and location of the antennas

Wavelength/frequency:
Long wavelength (low frequency), less loss
Short wavelength (high frequency), more loss

Different models to estimate path loss

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

17

Fading

Fast fading: rapid fluctuations in


amplitude, phase
Due to multi-path propagation resulting in
interference of multiple copies of the same
transmitted signal arriving at the receiver

Slow fading: the duration of the fading


may last for multiple seconds or minutes
Due to absorption of the transmission by
objects, for example in buildings

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

18

Interference

Adjacent channel interference:


interference caused by extraneous power
from a signal in an adjacent channel
Co-channel interference: due to weather
conditions, wireless communications systems
(radio, TV, etc.) in different locations that
share common channels can experience cochannel interference

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

19

Thermal Noise

Noise generated by the thermal agitation


of the charge carriers (the electrons)
inside an electrical conductor in
equilibrium, which happens regardless of
any applied voltage
Causes errors in reception (digital) or
degradation of quality (analog)
Effectively limits transmission range
when transmitting signal strength falls
below noise floor

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

20

Noise Limits Transmitting Distance

Signal to Noise Ratio


(SNR)

Short range transmission (low path loss)

=
High

Long range transmission (high path loss)


=

Low
Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

21

Transmission Rate Constraints

Nyquists Theorem: specifies the


maximum data rate possible on a
channel
C = 2 * B * log2 L (bits/sec)

B = bandwidth of the channel


C = maximum channel capacity
L = number of discrete signal levels/voltage

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

22

Transmission Rate Constraints

Shannons Theorem: specifies the


maximum data rate possible in a noisy
channel
C = B * log2 (1+ S/N) (bits/sec)

B = bandwidth of the channel


C = maximum channel capacity
S = signal power
N = noise power
SNR = 10 log10 (S/N)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

23

Multiple Access Techniques


Multiple access techniques: methods that determine
how the medium is accessed such the channel is
shared between multiple participants

Transmission medium is broadcast


If everybody sends, then communication is not
meaningful, garbage
Multiple access such that:
maximize message throughput
minimize mean waiting time

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

24

Multiple Access Methods

Three domains in which users can be


separated:
frequency, time, and space

Frequency division multiple access (FDMA)


Time division multiple access (TDMA)
Code division multiple access (CDMA)
Space division multiple access (SDMA)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

25

FDMA
Users are separated in frequency domain

Each station has its own frequency band,


separated by guard bands to eliminate interchannel interference
Receivers tune to the right frequency
Number of frequencies is limited
Best suited for analog links
Main drawback is under-utilization of the
frequency spectrum

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

26

TDMA
Users transmit data on same frequency, but at
different times (separated in time)

Relies on time synchronization


Users can be given different amounts of
bandwidth
Users can use idle times to determine best
base station
Synchronization overhead
Problems with multipath interference on
wireless links

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

27

CDMA
Users separated both by time and frequency

Send at a different frequency at each time slot


(frequency hopping)
Convert a single bit to a code (direct sequence),
receiver can decipher bit by inverse process
Difficult to spy
No need for all stations to synchronize
All cells can use all frequencies
Increased complexity

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

28

Spread Spectrum

Enables a signal to be transmitted across a


frequency band, much wider than the minimum
bandwidth required by the information signal
Transmitter spreads the energy across a
number of frequencies
Benefits: privacy, decreased narrowband
interference, increased signal capacity
In North America FCC waveband is divided
into 75 hopping channels, power transmission
limited to 1 watt on each channel

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

29

Frequency Hoping Spread Spectrum (FHSS)

Transmitter hops between available


frequencies according to a predefined
algorithm, which can be either random or
preplanned
Transmitter operates in synchronization with
the receiver
Large number of frequencies used

Results in a system that is quite resistant to


jamming

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

30

FHSS: Details

Signal hops from frequency to frequency at


fixed intervals
Transmitter operates in one channel at a time
Bits are transmitted using some encoding
scheme
At each successive interval, a new carrier
frequency is selected
Receiver, hopping between frequencies in
synchronization with transmitter, picks up
message

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

31

FHSS Example

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

32

Direct Sequence Spread Spectrum (DSSS)

Each bit in original signal is represented


by multiple bits in the transmitted signal
Data is chopped in small pieces and
spread across the frequency domain
Performance of DSSS is usually better
and more reliable

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

33

DSSS Example

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

34

DSSS vs FHSS

FHSS transmits data, using smaller


blocks of data, that are spread out over
many switching channels.
If two WLANs exist in the same area,
one of the two networks using FHSS,
FHSS will take precedence over the
DSSS-based network, which will only be
able to transport data correctly when the
FHSS-nodes are in idle mode.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

35

Space Division Multiple Access


Users are separated in space domain

Several users use the same frequency & time


slot (in TDMA)
Each user is separated by the smart antenna
by using its unique spatial location
Different areas can be served using the same
frequency
Increased co-channel interference from
adjacent co-channel cells

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

36

OSI/ISO Model

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

37

Summary : Wireless Characteristics:

Higher error rate


Higher signaling overhead

Constrained information
transmission
Distance vs data rates

Shared channel
Coordination

Open environment
Eavesdropping

Limited coverage
Cooperation

Cristina Nita-Rotaru
Lecture 1/ WS Milano Summer 2011

38

Wireless Networks:
Architectures and
Applications

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

39

Spectrum Allocation

Licensed: frequencies auctioned to the


highest bidder
Unlicensed: frequencies not allocated at all,
for example ISM (industrial, scientific and
medical)
Public bands

Name

900 Mhz

2.4 Ghz

5 Ghz

Range

902 - 928

2.4 - 2.4835

5.15 - 5.35

Bandwidth

26 Mhz

83.5 Mhz

200 Mhz

Wavelength .33m / 13.1 .125m / 4.9


Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

.06 m / 2.4
40

Spectrum Regulators

Federal Communications Commission (FCC)


regulates interstate and international
communications by radio, television, wire, satellite
and cable in US
Established by the Communications Act of 1934
and operates as an independent U.S. government
agency overseen by Congress.
Europe, on a country basis
Romania: National Regulatory Authority for
Communications and Information Technology
Italia: Ministro delle comunicazioni and
Autorit per le garanzie nelle comunicazioni.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

41

IEEE Standards
Slide by Omid Fatemieh (From Spring 2010 cs463 at UIUC)

WRAN

IEEE 802.22

Range

IEEE 802.20
WMAN

WiMax
IEEE 802.16

WLAN
ZigBee
802.15.4
Bluetooth
15.4c
802.15.1

WPAN
0.01
Cristina Nita-Rotaru
[Heile06]

0.1

WiFi
802.11

1
10
Data Rate (Mbps)
Lecture 1/ WS Milano Summer 2011

802.15.3
802.15.3c
100

1000
42

Wireless Networks

Cellular Networks
WMAN WiMAX
MANETs
WMNs
Sensor Networks
WPAN - Bluetooth
RFID
VANETs

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

43

Cellular Networks

Architecture: Wireless phones


communicate with an infrastructure
consisting of base stations
Uses licensed spectrum
Communication between base stations is wired
Frequency reused by dividing the area covered
by a cellular network in cells
Wide coverage
Large number of users

Applications: voice, text, data, video,


internet, payment

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

44

Wireless Local Area Networks (WLANs)

Provides increased bandwidth (up to 11Mb


for 802.11b and up to 54Mb for 802.11g)
Uses unlicensed spectrum to provide access
to a local network
Infrastructure mode: Fixed access point
connected to the wired infrastructure,
mobile stations communicating wireless
with the access point
Ad hoc mode: mobile stations
communicate with each other via wireless

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

45

Mobile Ad Hoc Networks (MANETs)

Architecture: Mobile nodes communicate via


multi-hop with each other
Requires cooperation
Network is self-configuring
Characteristics:
Easy to deploy, do not require fixed
infrastructure
Network topology may change rapidly and
unpredictably

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

46

Wireless Mesh Networks (WMNs)

Architecture: Set of fixed wireless routers


that form a wireless backbone and a set of
wireless clients. A WMN can be integrated
with other types of networks such as wired
(Internet), cellular or sensors networks via
a gateway
Applications: community and
neighborhood networking, broadband home
networking, enterprise networking, health
and medical systems, surveillance systems

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

47

WiMAX Networks

Architecture:
Long range system, licensed (2.3 GHz, 2.5 GHz and
3.5 GHz) or unlicensed spectrum
Uses a QoS mechanism based on connections
between the base station and the user device

Applications:
Mobile broadband connectivity across cities and
countries through a variety of devices.
Wireless alternative to cable and DSL for "last mile"
broadband access

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

48

Sensor Networks

Architecture:
Low cost small devices, able to sense the
environment (temperature, light, humidity),
report sensed data using wireless
communication
A large number of sensors (static or mobile),
distributed in an ad hoc manner over an area
Nodes cooperate: communicate via multi-hop
wireless communication, some nodes
aggregate data
Unlicensed spectrum
Applications: battlefield surveillance, medical
monitoring, biological detection, habitant
monitoring, home security, disaster recovery

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

49

SCADA

Architecture: (Supervisory Control and Data


Acquisition Type of Industrial Control System)
Human-Machine Interface (HMI)
Remote Terminal Units (RTUs): many are
wireless today
Programmable Logic Controllers (PLCs)
Devices hardened with respect to environmental
and physical threats
Applications: industrial control systems: computer
systems that monitor and control industrial,
infrastructure, or facility-based processes

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

50

Wireless Personal Area Networks: Bluetooth

Architecture:
Connect and exchange information in short
communication range of 1, 10 and 100 meters
Uses unlicensed spectrum

Applications:
PDAs, mobile phones, laptops, PCs, printers, digital
cameras and video game consoles
Cell phone and a hands free headset or car kit
PC input and output devices (mouse, keyboard and
printer)
Test equipment, GPS receivers and medical
equipment.
Remote controls where infrared was traditionally used

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

51

Vehicular Networks (VANETs)

Architecture: Cars have onboard


equipment equipped with GPS
Integrated with the safety belts, steering to prevent
skidding, and warning alerts
Can communicate with the roadside infrastructure

Applications: Congestion detection, collision


alert, toll collection, deceleration warning,
road hazard warning, electronic payments,
monitor traffic, send updates, and switch
traffic signals

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

52

VANETs Status

US: Spectrum allocated by FCC in 1999, existent standard:


Radio standard for Dedicated Short-Range Communications
(DSRC), based on an extension of 802.11, 802.11p
US Department of Transportation Initiative VII: Vehicle
Infrastructure Integration initiative will work toward deployment
of advanced vehicle-vehicle and vehicle-infrastructure
communications that could keep vehicles from leaving the road
and enhance their safe movement through intersections.
Europe in August 2008 the European Telecommunications
Standards Institute ETSI has allocated 30 MHz of spectrum in
the 5.9GHz band

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

53

Radio-Frequency Identification (RFID)

Architecture:
Wireless communication between a reader and
an electronic tag attached to an object
RFID tag: microchip + RF antenna, can be active
or passive, stores few hundred bits

Applications: identification and tracking,


electronic tickets for public transport systems,
access control to building, automated toll-payment
transponders, anti-theft systems for cars, passports

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

54

Next Generation of
GPS

Air Traffic Control


GPS

UAT/
1090
Squitter

HF/
Satcom
UAT

IP VPN
Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

55

Global Positioning System (GPS)

Architecture:
24 satellites that orbit the
Earth in very precise orbits
twice a day and emit
signals

Applications:
Position and coordinates
Travel progress reports
Accurate time
measurement

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

56

Galileo

Global navigation satellite system


currently being built by the European
Union.
Should become operational in
2013-2014.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

57

Wireless Networks: Summary so Far

Wireless communication is
omnipresent today operating in
licensed and unlicensed spectrum
Architectures:
Centralized
Peer to peer

Communication:
One-hop
Multi-hop

Devices: different computational power


and physical accessibility
Mobility:
Fixed node
Mobile nodes

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

58

Security Goals, Attacks,


Adversarial Models

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

59

Information Security

Security attacks: Any action that compromises


the security of information.
Security mechanism: A mechanism that is
designed to detect, prevent, or recover from a
security attack.
Security service: A service that enhances the
security of data processing systems and
information transfers. A security service makes
use of one or more security mechanisms.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

60

Security Services (or Goals)

1) Confidentiality: information is available for


reading only to authorized parties.
Example: Alice sends a message to Bob, only Alice and
Bob can understand the content of the message.

2) Authentication:
Data source authentication: the data is coming from an
authorized party.
Example: Alice receives a message from Bob. This
service ensures that the message is from Bob and not
from Carl.
Entity authentication: the entity is who it says it is.
Example: When Alice tries to obtain access to her
bank account, an authentication operation is performed
to ensure that Alice asks for the information.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

61

Security Services (2)

3) Integrity: detect if data was modified, from the


source to the destination.
Example: Alice sends an email to Bob. Carl
intercepts the message and modifies it. Data
integrity allows for Bob to detect that the message
was modified on the way from Alice to him.
4) Non-repudiation: neither the sender, nor the
receiver of a message are able to deny the
transmission.
Example: Alice sends Bob a contract, signed. The
non-repudiation service ensures that Alice can not
claim that the signature was produced by
somebody else.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

62

Security Services (3)

5) Access control: only authorized


parties can use specific resources.
Example: Alice wants to print a document, she must
be authorized to get that document and to use the
printer.

6) Availability: resources available to


authorized parties.
Example: A web site might become unavailable if
the server crashes, or is bombarded with requests.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

63

Security Attacks

Passive: the attacker does not modify the data,


only monitors the communication. It threatens
confidentiality.
Example: listen to the communication between
Alice and Bob, and if its encrypted try to decrypt it.
Active: the attacker is actively involved in deleting,
adding or modifying data. It threatens all security
services.
Example: Alice sends Bob a message: meet me
today at 5, Carl intercepts the message and
modifies it meet me tomorrow at 5, and then
sends it to Bob.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

64

Security Attacks: Examples

Interruption

Interception

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

65

Security Attacks: Examples

Modification

Fabrication (injection)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

66

Security Mechanisms

Cryptography: protect data by performing


operations on the data (for example encrypt data).
Software: access limitations to in a database, in
operating system protect each user from other
users, networking: firewall.
Hardware: use smartcards and trusted computing
for authentication.
Policies: define who has access to what resources.
Physical security: control who has
physical access to devices storing data.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

67

What is Cryptography About?

Constructing and analyzing protocols which


enable parties to achieve security objectives,
overcoming the influence of adversaries.
Note: a protocol (or a scheme) is a suite of
algorithms that tell each party what to do
Attack model: assumptions about the resources and
actions available to the adversary

How to devise and analyze protocols


understand the threats posed by the adversaries
and the secueity objectives (goals)
think as an adversary

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

68

Actually

Cryptography: the study of


mathematical techniques related to
aspects of providing information
security services (construct).
Cryptanalysis: the study of
mathematical techniques for
attempting to defeat information
security services (break).
Cryptology: the study of
cryptography and cryptanalysis
(both).

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

69

Basic Terminology in Cryptography

cryptography
cryptanalysis
cryptology
plaintexts
ciphertexts
keys
encryption
decryption

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

70

Secret-key vs. Public-key Cryptography

Secret-key cryptography (a.k.a. symmetric


cryptography)
encryption & decryption use the same key
key must be kept secret
key distribution is very difficult

Public-key cryptography (a.k.a. asymmetric


cryptography)
encryption key different from decryption key
cannot derive decryption key from encryption key
higher cost than symmetric cryptography

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

71

How Do You Know a Cipher is Secure?

Show that under the considered attack


model, security goals are NOT achieved
(break it)
Show that under the considered attack
model, security goals ARE achieved
(evaluate/prove)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

72

Breaking Ciphers

There are different methods


of breaking a cipher, depending on:
the type of information available to
the attacker
the interaction with the cipher
machine
the computational power available to
the attacker

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

73

Breaking Ciphers

Ciphertext-only attack:
The cryptanalyst knows only the
ciphertext. Sometimes the language of the
plaintext and the cipher are also known.
The goal is to find the plaintext and the key.

NOTE: any encryption scheme


vulnerable to this type of attack is
considered to be completely insecure.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

74

Breaking Ciphers (2)

Known-plaintext attack:
The cryptanalyst knows one

or
several pairs of ciphertext
and the corresponding
plaintext.

The goal is to find the key used to


encrypt these messages or a way
to decrypt any new messages that
use that key.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

75

Breaking Ciphers (3)

Chosen-plaintext attack
The cryptanalyst can choose a number of
messages and obtain the ciphertexts
for them
The goal is to deduce the key used in the other
encrypted messages or decrypt any new messages
using that key.

It can be adaptive, the choice


of plaintext depends on the
ciphertext received from
previous requests.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

76

Breaking Ciphers (4)

Chosen-ciphertext attack
Similar to the chosen-plaintext attack,
but the cryptanalyst can choose a
number of ciphertexts and obtain
the plaintexts.
It can also be adaptive The choice of
ciphertext may depend on the plaintext
received from previous requests.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

77

How Do You Know a Cipher is Secure?

Show that under the considered attack


model, security goals are NOT achieved
(break it)
Show that under the considered attack
model, security goals ARE achieved
(evaluate/prove)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

78

Models for Evaluating Security

Unconditional security
The adversary has unlimited computational
resources. Analysis is made by using probability
theory. Perfect secrecy: observation of the
ciphertext provides no information to an
adversary.
Complexity-theoretic security
The adversary is assumed to have polynomial
computational power. The analysis uses
complexity theory; Polynomial attacks although
feasible, in practice can be computationally
infeasible.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

79

Models for Evaluating Security (2)

Provable security
Proof of security relies on the difficulty of solving a
well-known and supposedly difficult problem
(example: computation of discrete logarithms).
Computational security (practical security)
Measures the amount of computational effort required
to defeat a system. Sometimes related to the hard
problems, but no proof of equivalence is known.
Ad hoc security (heuristic security)
Variety of convincing arguments that every
successful attack requires more resources than the
ones available to an attacker. Unforeseen attacks
remain a threat.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

80

One-Time Pad
Key is chosen randomly
Plaintext X = (x1 x2 xn)
Key
K = (k1 k2 kn)
Ciphertext Y = (y1 y2 yn)

ek(X) = (x1+k1 x2+k2 xn+kn) mod m


dk(Y) = (y1- k1 y2-k2 yn-kn) mod m

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

81

Shannon (Information-Theoretic) Security

Basic Idea: Ciphertext should provide no


information about Plaintext
Such a scheme has perfect secrecy
One-time pad has perfect secrecy if
Key-length msg-length
Key is random
Key is used only once

Result due to Shannon, 1949.


C. E. Shannon, Communication Theory of Secrecy Systems,
Bell System Technical Journal, vol.28-4, pp 656--715, 1949.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

82

Summary so Far

Cryptographic protocols are an


important tool in ensuring
security
Security goals: confidentiality
integrity, authentication
There are very few protocols for
which we can prove security
OTP has perfect secrecy under
some conditions
Ciphers that are vulnerable to
ciphertext-only attacks are
completely insecure

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

83

Stream Ciphers and Block


Ciphers

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

84

Stream Ciphers

OTP not practical for most applications, key needs to be


random, used only once and as long as the message
OPT: a key is a random bit string of length n
Stream ciphers:
Idea: replace rand by pseudo rand
Use Pseudo Random Number Generator
PRNG: {0,1}s {0,1}n
expand a short (e.g., 128-bit) random seed into a
long (e.g., 106 bit) string that looks random
Secret key is the seed
Eseed[M] = M PRNG(seed)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

85

Properties of Stream Ciphers

Do not have perfect secrecy


security depends on PRNG
PRNG must be unpredictable
given consecutive sequence of bits output
(but not seed), next bit must be hard to
predict
Typical stream ciphers are very fast
Used in many places, often incorrectly
SSL( RC4), DVD (LFSR), WEP (RC4), etc.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

86

Fundamental Weaknesses of Stream Ciphers

If the same keystream is used


twice ever, then easy to break
Highly malleable
easy to change ciphertext so that
plaintext changes in predictable,
e.g., flip bits

Weaknesses exist even if the


PRNG is strong

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

87

Block Ciphers

Map n-bit plaintext blocks to n-bit


ciphertext blocks (n: block length).
For n-bit plaintext and ciphertext blocks
and a fixed key, the encryption function
is a bijection; E : Pn X K Cn s.t. for all
key k K, E(x, k) is an invertible
mapping written Ek(x).
The inverse mapping is the decryption
function, y = Dk(x) denotes the
decryption of plaintext x under k.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

88

Block Ciphers Features

Block size: in general larger block sizes


mean greater security.
Key size: larger key size means greater
security (larger key space).
Number of rounds: multiple rounds offer
increasing security.
Encryption modes: define how messages
larger than the block size are encrypted, very
important for the security of the encrypted
message.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

89

History of Data Encryption Standard (DES)

1967: Feistel at IBM


Lucifer: block size 128; key size 128 bit
1972: NBS asks for an encryption standard
1975: IBM developed DES (modification of Lucifer)
block size 64 bits; key size 56 bits
1975: NSA suggests modifications
1977: NBS adopts DES as encryption standard in
(FIPS 46-1, 46-2).
2001: NIST adopts Rijndael as replacement to DES.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

90

DES Features

Features:
Block size = 64 bits
Key size = 56 bits
Number of rounds = 16
16 intermediary keys, each 48
bits

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

91

Cryptanalysis of DES
Brute Force:
Known-Plaintext Attack
56 possible keys
Try all 2
Requires constant memory
Time-consuming
DES challenges: (RSA)

Cristina Nita-Rotaru

msg=the unknown message is :xxxxxxxx


CT= C1
| C2
| C3 | C4
1997 Internet search: 3 months
1998 EFF machine (costs $250K): 3 days
1999 Combined: 22 hours

Lecture 1/ WS Milano Summer 2011

92

Triple DES

Use three different keys


Encrypt: C = EK3 [ DK2 [ EK1 [P] ] ]
Decrypt: P = DK3 [ EK2 [ DK1 [C] ] ]

Key space is 56 x 3 = 168 bits


No known practical attack
against it.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

93

AES - Rijndael Features

Designed to be efficient in both hardware and


software across a variety of platforms.
Uses a variable block size, 128,192, 256-bits, key
size of 128-, 192-, or 256-bits.
128-bit round key used for each round (Can be precomputed and cached for future encryptions).
Note: AES uses a 128-bit block size.
Variable number of rounds (10, 12, 14):
10 if B = K = 128 bits
12 if either B or K is 192 and the other is 192
14 if either B or K is 256 bits

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

94

Rijandel Cryptanalysis
Resistant to linear and differential cryptanalysis

Academic break on weaker version of


the cipher, 9 rounds
Requires 2224 work and 285 chosen
related-key plaintexts.
Attack not practical.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

95

Encryption Modes: ECB

Message is broken into independent


blocks of block_size bits;
Electronic Code Book (ECB): each block
encrypted separately.
Encryption: ci = Ek(xi)
Decryption: xi = Dk(ci)

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

96

Properties of ECB

Deterministic: the same data block gets


encrypted the same way, reveals
patterns of data when a data block
repeats.
Malleable: reordering ciphertext results in
reordered plaintext.
Errors in one ciphertext block do not
propagate.
Usage: not recommended to encrypt
more than one block of data.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

97

Encryption Modes: CBC

Cipher Block Chaining (CBC): next


input depends upon previous output
Encryption: Ci= Ek (MiCi-1), with C0=IV
Decryption: Mi= Ci-1Dk(Ci), with C0=IV

IV

C0
Cristina Nita-Rotaru

M1

M2

M3

Ek

Ek

Ek

C1

C2

C3

Lecture 1/ WS Milano Summer 2011

98

Properties of CBC

Randomized encryption: repeated text gets mapped to


different encrypted data.
can be proven to be secure assuming that the block cipher has
desirable properties and that random IVs are used

A ciphertext block depends on all preceding plaintext blocks;


reorder affects decryption
Self-correcting: errors in one block propagate to two blocks
Sequential encryption: cannot use parallel hardware
Usage: chooses random IV and protects the integrity of IV
Observation: if Ci=Cj then Ek (MiCi-1) = Ek (MjCj-1); thus MiCi-1
= MjCj-1; thus Mi Mj = Ci-1 Cj-1

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

99

Use Block Ciphers to Construct Stream Ciphers

Cipher Feedback (CFB)


Output Feedback (OFB)
Counter Mode (CTR)
Common properties:
uses only the encryption function of the cipher
both for encryption and for decryption
malleable: possible to make predictable bit
changes

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

100

Encryption Modes: CFB

Cipher Feedback (CFB): the message is


XORed with the feedback of encrypting the
previous block
Decryption

Encryption
r-bit shift

r-bit shift

I1=IV

Ij

Ij

Oj
xj
Cristina Nita-Rotaru

cj

cj

Lecture 1/ WS Milano Summer 2011

Oj
xj
101

Properties of CFB

Randomized encryption
A ciphertext block depends on all preceding
plaintext blocks; reorder affects decryption
Errors propagate for several blocks after
the error, but the mode is selfsynchronizing (like CBC).
Decreased throughput.
Can vary the number of bits feed back, trading off
throughput for ease of use

Sequential encryption

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

102

Encryption Modes: OFB

Output feedback (OFB):


construct a PRNG using DES
y0=IV yi = Ek[yi-1]

Decryption

Encryption
Oj-1

Oj-1

I1=IV

Ij

Ij

I1=IV

Oj
xj
Cristina Nita-Rotaru

Oj
cj

cj

Lecture 1/ WS Milano Summer 2011

xj
103

Properties of OFB

Randomized encryption
Sequential encryption, but preprocessing possible
Error propagation limited
Subject to limitation of stream cipher

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

104

Encryption Modes:CTR

Counter Mode (CTR): Another way


to construct PRNG using DES
yi = Ek[counter+i]
Sender and receiver share: counter
(does not need to be secret) and the
secret key.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

105

Properties of CTR

Software and hardware efficiency: different


blocks can be encrypted in parallel.
Preprocessing: the encryption part can be done
offline and when the message is known, just do
the XOR.
Random access: decryption of a block can be
done in random order, very useful for hard-disk
encryption.
Messages of arbitrary length: ciphertext is the
same length with the plaintext (i.e., no IV).

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

106

Ideal Block Cipher

An ideal block cipher is a substitution


cipher from {0,1}n to {0,1}n
Also known as a random permutation
Each key determines one permutation on the
plaintext space
A random key is chosen

Why is this an ideal block cipher?


Known-plaintext, chosen plaintext, and
chosen ciphertext attacks are totally
ineffective

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

107

Security Goal of Block Cipher

Indistinguishable from an ideal block


cipher (i.e., a random permutation)
The best block cipher should be a
pseudo-random permutation (PRP)
For all existing block ciphers, if there is
no known attack, they are assumed to
be PRP for some suitable parameters.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

108

Block Cipher Modes Revisited

Suppose that the adversary knows that a


ciphertext results from one of two
possible plaintexts, the adversary should
not be able to tell that which one
plaintext is more likely to be the actual
one.
If a block cipher is a PRP, then using this
cipher under the CBC, CTR modes has
semantic security.

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

109

Summary so Far

Stream ciphers are faster than


block ciphers
Keystream reuse for stream
ciphers makes them insecure
Current standard is AES, no
known practical attacks against it
Security of block ciphers
depends on the encryption mode
Recommended encryption
modes CBC and CTR

Cristina Nita-Rotaru

Lecture 1/ WS Milano Summer 2011

110

Você também pode gostar