iLab

Modern cryptography for communications security

part 1

Benjamin Hof
[email protected]

Lehrstuhl für Netzarchitekturen und Netzdienste

Fakultät für Informatik
Technische Universität München

Cryptography – 18ss

1 / 34
Outline

Cryptography

Symmetric setting

2 / 34
Outline

Cryptography

Symmetric setting

3 / 34
Scope

Focus on:
I modern cryptography
I methods used in communications security

Based on: Introduction to modern cryptography, Katz and Lindell,

2nd edition, 2015.

4 / 34
Communication

by Melissa Elliott

https://twitter.com/0xabad1dea/status/400676797874208768

5 / 34
What we are concerned with

“Let’s meet up at 9!”

Alice Bob

6 / 34
What we are concerned with

“Let’s meet up at 9!”

Alice Bob

BfV

Roens/Wikipedia. CC-by-sa 2.0

6 / 34
What we are concerned with

“Let’s meet up at 9!”

Alice Bob

Eve

passive attack: eavesdropping

We want to provide confidentiality!

6 / 34
What we are concerned with

“This will not be on the exam!”

Alice Mallory Bob

active attack: message modification or forgery

We want to provide message authentication!

6 / 34
Limitations

I cryptography is typically bypassed, not broken

I not applied correctly
I not implemented correctly
I subverted

No protection of information about the communication.

I existence
I time
I extent
I partners

7 / 34
Kerckhoffs’ principle

Security should only depend on secrecy of the key, not the

secrecy of the system.
I key easier to keep secret
I change
I compatibility

No security by obscurity.
I scrutiny
I standards
I reverse engineering

8 / 34
Another principle as a side note

The system should be usable easily.

I Kerckhoffs actually postulated 6 principles
I this one got somewhat forgotten
I considered uncontroversial by Kerckhoffs
I starting to be rediscovered in design of secure applications and
libraries

Example
Signal, NaCl

9 / 34
What should secure encryption guarantee?

It should be impossible for the attacker to

10 / 34
What should secure encryption guarantee?

It should be impossible for the attacker to

I recover the key.
I recover the entire plaintext from the ciphertext.
I recover any character of the plaintext from the ciphertext.

10 / 34
What should secure encryption guarantee?

It should be impossible for the attacker to

I recover the key.
I recover the entire plaintext from the ciphertext.
I recover any character of the plaintext from the ciphertext.

Regardless of any information an attacker already has, a ciphertext

should leak no additional information about the underlying plaintext.

10 / 34
Modern cryptography

relies on
I formal definitions
I precisely defined assumptions
I mathematical proofs
Reductionist security arguments, the proofs, require to formulate
assumptions explicitly.

11 / 34
A definition of security

A scheme is secure, if any probabilistic polynomial time adversary

succeeds in breaking the scheme with at most negligible probability.

Negligible
For every polynomial p and for all sufficiently large values of n:
1
f (n) <
p(n)
1
e.g., f (n) = 2n

Church-Turing Hypothesis
We believe polynomial time models all computers.

12 / 34
Our goals

symmetric (secret-key) asymmetric (public-key)

I confidentiality I confidentiality
I authenticity I authenticity
(as in: message integrity) I key exchange

Something providing confidentiality generally makes no statement

whatsoever about authenticity.

13 / 34
Motivation

What does a perfectly encrypted message look like?

14 / 34
Uniform distribution

P : U → [0, 1]

X
P(x ) = 1
x ∈U

1
∀x ∈ U : P(x ) =
|U|

15 / 34
Randomness

I required to do any cryptography at all

I somewhat difficult to get in a computer (deterministic!)
I required to be cryptographically secure: indistiguishable from
truly random
I not provided in programming languages

Example
used to generate keys or other information unkown to any other
parties

16 / 34
Collecting unpredictable bits

I physical phenomena
I time between emission of particles during radioactive decay
I thermal noise from a semiconductor diode or resistor
I software-based
I elapsed time between keystrokes or mouse movement
I packet interarrival times

I attacker must not be able to guess/influence the collected

values

1. collect pool of high-entropy data

2. process into sequence of nearly independent and unbiased bits

17 / 34
Pseudo-random generator

G : {0, 1}s → {0, 1}n , ns

18 / 34
Outline

Cryptography

Symmetric setting

19 / 34
Symmetric encryption scheme

1. k ← Gen(1n ), security parameter 1n

2. c ← Enck (m), m ∈ {0, 1}∗
3. m := Deck (c)

I provide confidentiality
I definition of security: chosen-plaintext attack (CPA)
Cryptography uses theoretical attack games to analyze and
formalize security.

C: challenger, ← means non-deterministic,

A: adversary := means deterministic

20 / 34
The eavesdropping experiment

C A

k ← Gen(1n ) input 1n
The eavesdropping experiment

C A

k ← Gen(1n ) input 1n
m 0, m 1

b ← {0, 1}
c ← Enck (mb )
c

output b 0

I A succeeds, iff b = b 0

21 / 34
Discussion of the eavesdropping experiment

I |m0 | = |m1 |
I probabilistic polynomial time algorithms

I success probability should be 0.5 + negligible

I if so, Enc has indistinguishable encryptions in the presence of
an eavesdropper

22 / 34
Pseudorandom permutation

F : {0, 1}∗ × {0, 1}∗ → {0, 1}∗

I Fk (x ) and Fk−1 (y ) efficiently computable

I Fk be indistinguishable from uniform permutation
I adversary may have access to F −1
We can assume that all inputs and the output have the same length.

23 / 34
A block cipher

Example
I fixed key length and block length
I chop m into 128 bit blocks

m k

128 bit
AES

Does this function survive the eavesdropping experiment?

24 / 34
Chosen-plaintext attack
C A

k ← Gen(1n ) input 1n

25 / 34
Chosen-plaintext attack
C A

k ← Gen(1n ) input 1n

m
c ← Enck (m)
c

.. ..
. .

25 / 34
Chosen-plaintext attack
C A

k ← Gen(1n ) input 1n

m
c ← Enck (m)
c

.. ..
. .

m0, m1
b ← {0, 1}
Enck (m
b)

25 / 34
Chosen-plaintext attack
C A C (cont’d) A

k ← Gen(1n ) input 1n m

m c ← Enck (m)
c
c ← Enck (m)
c .. ..
. .
.. ..
. . output bit b 0
m0, m1
b ← {0, 1}
Enck (m
b)

25 / 34
Chosen-plaintext attack
C A C (cont’d) A

k ← Gen(1n ) input 1n m

m c ← Enck (m)
c
c ← Enck (m)
c .. ..
. .
.. ..
. . output bit b 0
m0, m1
b ← {0, 1}
Enck (m
b)

25 / 34
Discussion of CPA

I Enc is secure under chosen-plaintext attack

I again, messages must have same length
I multiple-use key
I non-deterministic (e. g. random initialization vector) or state
I block cipher requires operation mode, e. g.: counter (CTR),
output-feedback (OFB), . . .

26 / 34
Example constructions: counter mode
Example
I randomised AES counter mode (AES-CTR$)
I choose nonce r ← {0, 1}128 , key k ← {0, 1}128
I great if you have dedicated circuits for AES, else vulnerable to
timing attacks

r AES k r +1 AES k

m0 ⊕ m1 ⊕

c0 c1
···

complete ciphertext c := (r , c0 , c1 , · · · )
27 / 34
Example constructions: stream ciphers
Example
A modern stream cipher, fast in software:

96 bit nonce 32 bit initial counter 128 bit key

ChaCha

keystream
plaintext ⊕

ciphertext

28 / 34
Message authentication code (MAC)

1. k ← Gen(1n ), security parameter 1n

2. t ← Mack (m), m ∈ {0, 1}∗
3. b := Vrfyk (m, t)
b = 1 means valid, b = 0 invalid

I transmit hm, ti
I tag t is a short authenticator
I message authenticity ⇔ integrity
I detect tampering
I no protection against replay
I “existentially unforgeable”
I security definition: adaptive chosen-message attack

29 / 34
Adaptive chosen-message attack

C A
k ← Gen(1n )

m input 1n

t ← Mack (m) hm, ti

.. ..
. .
output hm0 , t 0 i

I let Q be the set of all queries m

I A succeeds, iff Vrfyk (m0 , t 0 ) = 1 and m0 ∈
/Q

30 / 34
Used in practice

Example
I HMAC based on hash functions
I CMAC based on cipher block chaining mode (CBC)
I authenticated encryption modes

31 / 34
Example: side-channel attack

How does tag verification work and how to implement tag

comparison correctly?

32 / 34
Recap: secret-key cryptography

I attacker power: probabilistic polynomial time

I confidentiality defined as IND-CPA:
encryption, e. g. AES-CTR$
I message authentication defined as existentially unforgeable
under adaptive chosen-message attack:
message authentication codes, e. g. HMAC-SHA2
I authenticated encryption modes

33 / 34
Combining confidentiality and authentication

I encrypt-then-authenticate is generally secure:

c ← Enc k1 (m), t ← Mac k2 (c)
transmit: hc, ti
I authenticated encryption is also a good choice:
e. g. offset codebook (OCB), Galois counter mode (GCM)
c, t ← AEAD enck (ad, m)
m := AEAD dec k (ad, c, t) or verification failure

34 / 34