Cryptography and Network Security

Bhaskaran Raman
Department of CSE, IIT Kanpur Reference: Whitfield Diffie and Martin E. Hellman, Privacy and Authentication: An Introduction to Cryptography , in Proc. IEEE, vol. 67, no.3, pp. 397 - 427, 1979

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Cryptography Fundamentals

Privacy versus Authentication:

Privacy: preventing third party from snooping Authentication: preventing impostering Guarantee that no third party has modified data Receiver can prove that only the sender originated the data

Two kinds of authentication:

Digital Signature E.g., for electronic transactions

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Cryptographic Privacy
Eavesdropper Sender P Encryption C=SK(P)

Network
Key:K

Decryption C=S1K(P)

Receiver

Encrypt before sending, decrypt on receiving

Terms: plain text and cipher text Should algorithm be secret?

Two components: key, and the algorithm

Yes, for military systems; no, for commercial systems

Key distribution must be secure

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Cryptographic Authentication
Eavesdropper P C' Encryption C=SK(P) P'

Sender

Network
Key:K

Decryption C'=S1K(P')

Receiver

The same system can also be used for authentication

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Cryptanalysis

Cryptanalysis: attacker tries to break the system

E.g., by guessing the plain text for a given cipher text Or, by guessing the cipher text for some plain text Cipher-text only attack Known plain-text attack Chosen plain-text attack Chosen text attack
FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Possible attacks:

Security Guarantees

Two possibilities:

Unconditional Computational security One-time tape How much security to have? Depends on cost-benefit analysis for attacker

Unconditional security: an example

Most systems have computational security

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Public-Key Systems

Shared-key ==> difficulties in key distribution

C(n,2) = O(n^2) keys Public component and a private component Two kinds:

Public key system

Public key distribution: establish shared key first Public key cryptography: use public/private keys in encryption/decryption

Public key cryptography can also be used for digital signatures

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Some Example Systems

Permuted alphabet (common puzzle)

Can be attacked using frequency analysis, patterns, digrams, trigrams Attack becomes difficult if alphabet size is large

Transposition Poly-alphabetic: periodic or running key Codes versus ciphering

Codes are stronger, and also achieve data compression

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Some Popular Systems

Private key systems:

DES, 3DES RSA: based on difficulty of factoring Galois-Field (GF) system: based on difficulty of finding logarithm Based on knapsack problem

Public key systems:

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Digital Encryption Standard (DES)

64bits

64bits Key

64bits

Plaintext

Ciphertext

R1

R2

R16

P1

Permutation,16roundsofidenticaloperation,inversepermutation

Li1 Eachroundusesa different48bitkey Ki(fromK)anda combinerfunctionF Li1

Ri1
F

Ki

+ Ri1

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Triple-DES (3DES)

DES can be broken with 2^55 tries:

4500 years on an Alpha workstation But only 6 months with 9000 Alphas Use DES thrice, with 3 separate keys, or with two keys (K1 first, then K2, then K1 again)

Triple-DES:

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Rivest, Shamir, Adleman (RSA) Public-Key Crypto-System

Based on the fact that finding large (e.g. 100 digit) prime numbers is easy, but factoring the product of two such numbers appears computationally infeasible Choose very large prime numbers P and Q

N=PxQ N is public; P, Q are secret

Euler totient: Phi(N) = (P-1)(Q-1) = Number of integers less than N & relatively prime to N
FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

RSA (continued)

Next, choose E in [2, Phi(N)-1], E is public A message is represented as a sequence M1, M2, M3..., where each M in [0, N-1] Encryption: C = ME mod N Using the secret Phi(N), A can compute D such that ED = 1 mod Phi(N) ED = k x Phi(N) + 1 Then, for any X < N, Xk x Phi(N)+1 = X mod N
FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

RSA (Continued)

Decryption: CD = MED = Mk x Phi(N)+1 = M mod N Example: Choose P = 17, Q = 31

N = 527, Phi(N) = 480 Choose E = 7, then D = 343 If M = 2, Encryption: C = 128 Decryption: D = CD mod N = 128343 mod 527 = 2

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Taxonomy of Ciphers

Block ciphers: divide plain text into blocks and encrypt each independently Properties required:

No bit of plain text should appear directly in cipher text Changing even one bit in plain text should result in huge (50%) change in cipher text Exact opposite of properties required for systematic error correction codes

Stream cipher: encryption depends on current state

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Key Management

Keys need to be generated periodically

New users Some keys may be compromised

Addressing the O(n^2) problem with key distribution

Link encryption Key Distribution Centre (KDC): all eggs in one basket Multiple KDCs: better security

Key management easier in public key cryptography

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005

Some Non-Crypto Attacks

Man-in-the-middle attack: play a trick by being in the middle Traffic analysis:

Can learn information by just looking at presence/absence of traffic, or its volume Can be countered using data padding To counter: need to verify timeliness of message from sender while authenticating Beware of issues of time synchronization

Playback or replay attacks:

FundamentalsofWiredandWirelessNetworks,KameswariChebroluandBhaskaranRaman,0913May2005