Brief Overview of
Cryptography
Outline
cryptographic primitives
– symmetric key ciphers
• block ciphers
• stream ciphers
– asymmetric key ciphers
– cryptographic hash functions
protocol primitives
– block cipher operation modes – “enveloping”
– message authentication codes – digital signatures
key management protocols
– session key establishment with symmetric and asymmetric key techniques
– Diffie-Hellman key exchange and the man-in-the-middle attack
– public key certification
C ry p to gr a p hi c pr im iti ve s
E E D D
x
plaintext
k
encryption key
k’
decryption key Ek(x)
ciphertext Dk’ (Ek(x)) = x
attacker
Operational model of encryption
Kerckhoff’s assumption:
– attacker knows E and D
– attacker doesn’t know the (decryption) key
attacker’s goal:
– to systematically recover plaintext from ciphertext – to deduce the (decryption) key
attack models:
– ciphertext-only – known-plaintext
– (adaptive) chosen-plaintext – (adaptive) chosen-ciphertext
block ciphers
block cipher block cipher
plaintext ciphertext
padding key
Symmetric key encryption
it is easy to compute k from k’ (and vice versa)
often k = k’
two main types: stream ciphers and block ciphers
pseudo-random bit stream generator
pseudo-random bit stream generator
...
plaintext+ ...
ciphertextstream ciphers
seedp to gr a p hi c pr im iti ve s
One-time pad – theoretical vs.
practical security
one-time pad
– a stream cipher where the key stream is a true random bit stream
– unconditionally secure (Shannon, 1949)
– however, the key must be as long as the plaintext to be encrypted
practical ciphers
– use much shorter keys
– are not unconditionally secure, but computationally infeasible to break
– however, proving that a cipher is computationally secure is not easy
• not enough to consider brute force attacks (key size) only
• a cipher may be broken due to weaknesses in its (algebraic) structure
– no proofs of security exist for many ciphers used in practice – if a proof exists, it usually relies on assumptions that are
widely believed to be true (such as P NP)
C ry p to gr a p hi c pr im iti ve s
DES – Data Encryption Standard
input size: 64, output size: 64, key size: 56
16 rounds
Feistel structure
– F need not be invertible – decryption is the same
as encryption with
reversed key schedule (hardware
implementation!)
Initial Permutation Initial Permutation
FF
+
FF
+
FF
+
FF
+
…
Initial Permutation-1
(64)
(32) (32)
(48)
(48)
(48)
(48)
Key Scheduler
(56)
K K1
K2
K16 K3 X
ry p to gr a p hi c pr im iti ve s
DES round function F
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
S1S1 S2S2 S3S3 S4S4 S5S5 S6S6 S7S7 S8S8
PP
Si – substitution box (S-box) (look-up table)
P – permutation box (P-box)
C ry p to gr a p hi c pr im iti ve s
DES key scheduler
Permuted Choice 1 Permuted Choice 1
Permuted Choice 2 Permuted Choice 2 Left shift(s) Left shift(s)
Permuted Choice 2 Permuted Choice 2 Left shift(s) Left shift(s)
…
(28) (56)
K
(28)
(28) (28)
(48)
(48)
K1
K2
each key bit is used in around 14 out of 16 rounds
p to gr a p hi c pr im iti ve s
AES – Advanced Encryption Standard
NIST selected Rijndael (designed by Joan
Daemen and Vincent Rijmen) as a successor of DES (3DES) in November 2001
Rijndael parameters
– key size 128 192 256
– input/output size 128 128 128 – number of rounds 10 12 14 – round key size 128 128 128
not Feistel structure
decryption algorithm is different from encryption algorithm (optimized for encryption)
single 8 bit to 8 bit S-box
key injection (bitwise XOR)
C ry p to gr a p hi c pr im iti ve s
General structure of Rijndael encryption/decryption
p to gr a p hi c pr im iti ve s
add round key substitute bytes
shift rows mix columns add round key
substitute bytes shift rows mix columns add round key substitute bytes
shift rows add round key
plaintext
add round key inverse subs bytes
inverse shift rows inverse mix columns
add round key inverse subs bytes
inverse shift rows
inverse mix columns add round key inverse subs bytes
inverse shift rows add round key
plaintext w[0..3]
w[4..7]
w[36..39]
w[40..43]
expanded key
round 1round 9round 10 round 1round 9round 10
Rijndael – Shift row and mix column
C ry p to gr a p hi c pr im iti ve s
s00 s10 s20 s30
s01 s11 s21 s31
s02 s12 s22 s32
s03 s13 s23 s33
s00 s11 s22 s33
s01 s12 s23 s30
s02 s13 s20 s31
s03 s10 s21 s32
LROT1 LROT2 LROT3
shift row
s00 s10 s20 s30
s01 s11 s21 s31
s02 s12 s22 s32
s03 s13 s23 s33
s’00 s’10 s’20 s’30
s’01 s’11 s’21 s’31
s’02 s’12 s’22 s’32
s’03 s’13 s’23 s’33
mix column
2 3 1 1 1 2 3 1 1 1 2 3 3 1 1 2
x =
multiplications and additions are performed over GF(28)
Rijndael – Key expansion
p to gr a p hi c pr im iti ve s
k0 k1 k2 k3
k4 k5 k6 k7
k8 k9 k10 k11
k12 k13 k14 k15
w0 w1 w2 w3
w4 w5 w6 w7
+
gg
+ + +
w8 w9 w10 w11
+
gg
+ + +
… function g
- rotate word - substitute bytes
- XOR with round constant
RC4 stream cipher
initialization (input: a seed K of keylen bytes)
for i = 0 to 255 do S[i] = i;
T[i] = K[i mod keylen];
initial permutation
j = 0;
for i = 0 to 255 do
j = (j + S[i] + T[i]) mod 256;
swap(S[i], S[j]);
stream generation (output: a stream of pseudo- random bytes)
i, j = 0;
while true
i = (i + 1) mod 256;
j = (j + S[i]) mod 256;
swap(S[i], S[j]);
t = (S[i] + S[j]) mod 256;
output S[t];
C ry p to gr a p hi c pr im iti ve s
Asymmetric key encryption
breakthrough of Diffie and Hellman, 1976
it is hard (computationally infeasible) to compute k’ from k
k can be made public (public-key cryptography)
E E D D
x
plaintext
k
encryption key
k’
decryption key Ek(x)
ciphertext Dk’ (Ek(x)) = x
attacker
p to gr a p hi c pr im iti ve s
RSA (Rivest, Shamir, Adleman, 1978)
basis
– computing xe mod n is easy but x1/e mod n is hard (n is composite)
– intractability of integer factoring
key generation
– select p, q large primes (about 500 bits each) – n = pq, (n) = (p-1)(q-1)
– select e such that 1 < e < (n) and gcd(e, (n)) = 1
– compute d such that ed mod (n) = 1 (this is easy if p and q are known)
– public key is (e, n) – private key is d
encryption
c = me mod n where m < n is the message
decryption
cd mod n = m
C ry p to gr a p hi c pr im iti ve s
Proof of RSA decryption
Fermat’s theorem
Let r be a prime. If gcd(a, r) = 1, then a
r-1mod r = 1.
Euler’s generalization
For every a and n where gcd(a, n) = 1, a
(n)mod n
= 1.
RSA decryption
c
dmod n
= (m
emod n)
dmod n
= m
edmod n
= m
k(n)+1mod n
= m*(m
(n))
kmod n
= m*(m
(n)mod n)
kmod n
if gcd(m, n) = 1= m mod n = m
p to gr a p hi c pr im iti ve s
Proof of RSA decryption cont’d
RSA decryption if gcd(m, n) > 1
– either p|m or q|m
– assume without loss of generality that p|m – note that in this case, q|m cannot hold since
otherwise m pq = n
– this means that gcd(m, q) = 1 cd mod p = med mod p = 0
cd mod q = med mod q = mk(p-1)(q-1)+1 mod q = m*(m (q-
1)) k(p-1) mod q =
m*(m (q-1) mod q) k(p-1) mod q = m mod q
p,q|(cd – m)
cd – m = spq = sn
cd = sn + m
cd mod n = m mod n = m
Cryptographic hash functions
requirements
– one-way: given a hash value y, it is computationally infeasible to find a message x such that h(x) = y
– weak collision resistance: given a message x, it is computationally infeasible to find another message x’ such that h(x) = h(x’)
– (strong) collision resistance: it is computationally infeasible to find two messages x and x’ such that h(x) = h(x’)
message of arbitrary length
fix length
message digest / hash value / fingerprint
p to gr a p hi c pr im iti ve s
hash function hash function
How long should a hash value be?
birthday paradox
– P(n, k) = Pr{ there exists at least one duplicate among k items where
each item can take on one of n equally likely values}
– P(n, k) > 1 – exp( -k*(k-1)/2n )
– Q: What value of k is needed such that P(n, k) > 0.5 ? – A: k should approximately be n0.5
– e.g., P(365, 23) > 0.5
birthday paradox applied to hash function h
– n is the number of possible hash values
– one can find a collision among n0.5 messages with probability greater than 0.5
– if output size of h is 64 bits, then n0.5 is 232 too small – output size should be at least 128 but 160 is even
better
C ry p to gr a p hi c pr im iti ve s
General structure of hash functions
if the compression function f is collision resistant, then so is the iterated hash function (Merkle and Damgard, 1989)
if necessary, the final block is padded to b bits
the final block also includes the total length of the input (this makes the job of an attacker
more difficult)
ff
X1
CV0
(b)
(n) (n)
CV1 ff
X2
(b)
(n)
CV2 ff
X3
(b)
(n)
CV3 CVL-1 ff
XL
(b)
(n) h(X)
…
p to gr a p hi c pr im iti ve s
SHA1 – Secure Hash Algorithm
output size (n): 160 bits
input block size (b): 512 bits
padding is always used
CV
0A = 67 45 23 01 B = EF CD AB 89 C = 98 BA DC FE D = 10 32 54 76 E = C3 D2 E1 F0
C ry p to gr a p hi c pr im iti ve s
10000000 … 00000 length
512 bits
64 bits
last input block
SHA1 compression function f
ry p to gr a p hi c pr im iti ve s
f[0..19], K[0..19], W[0..19]
20 steps
f[0..19], K[0..19], W[0..19]
20 steps
f[20..39], K[20..39], W[20..39]
20 steps
f[20..39], K[20..39], W[20..39]
20 steps
f[40..59], K[40..59], W[40..59]
20 steps
f[40..59], K[40..59], W[40..59]
20 steps
f[60..79], K[60..79], W[60..79]
20 steps
f[60..79], K[60..79], W[60..79]
20 steps
+ + + + +
A B C D E
A B C D E
A B C D E
CVi - 1
(5 x 32 = 160)
Xi
(512)
mod 232 additions
SHA1 compression function f cont’d
C ry p to gr a p hi c pr im iti ve s
LROT5 LROT5
+
LROT30 LROT30
f[t]f[t]
+ + +
A B C D E
A B C D E
W[t]
K[t]
mod 232 additions
SHA1 compression function f cont’d
f[t](B, C, D)
t = 0..19 f[t](B, C, D) = (B C) (B D) t = 20..39 f[t](B, C, D) = B C D
t = 40..59 f[t](B, C, D) = (B C) (B D) (C D) t = 60..79 f[t](B, C, D) = B C D
W[t]
W[0..15] = Xi
t = 16..79 W[t] = LROT1(W[t-16] W[t-14] W[t- 8] W[t-3])
K[t]
t = 0..19 K[t] = 5A 82 79 99 [230 x 21/2] t = 20..39 K[t] = 6E D9 EB A1 [230 x 31/2] t = 40..59 K[t] = 8F 1B BC DC [230 x 51/2] t = 60..79 K[t] = CA 62 C1 D6 [230 x 101/2]
p to gr a p hi c pr im iti ve s
Block cipher operation modes – ECB
Electronic Codebook (ECB)
– encrypt
– decrypt
P ro to co l p ri m iti ve s
EE
P1
C1
K EE
P2
C2
K EE
PN
CN
… K
DD
C1
P1
K DD
C2
P2
K DD
CN
PN K
…
Block cipher operation modes – CBC
Cipher Block Chaining (CBC)
– encrypt
– decrypt
EE
P1
C1 K
+
EE
P2
C2 K
+
EE
P3
C3 K
+
EE
PN
CN-1 K
IV CN-1 +
…
DD
C1 K
IV +
DD
C2 K
+
DD
C3 K
+
DD
CN K
+ CN-1
to co l p ri m iti ve s
Block cipher operation modes – CFB
Cipher Feedback (CFB)
– encrypt – decrypt
EE
Pi Ci
K
+
shift register (n)
(n)
select s bits select s bits
(n)
(s)
(s) (s)
(s)
initialized with IV
EE
Ci Pi
K
+
shift register (n)
(n)
select s bits select s bits
(n)
(s)
(s) (s)
(s)
initialized with IV
P ro to co l p ri m iti ve s
Block cipher operation modes – OFB
Output Feedback (OFB)
– encrypt – decrypt
EE
Pi Ci
K
+
shift register (n)
(n)
select s bits select s bits
(n)
(s)
(s) (s)
(s)
initialized with IV
EE
Ci Pi
K
+
shift register (n)
(n)
select s bits select s bits
(n)
(s)
(s) (s)
(s)
initialized with IV
to co l p ri m iti ve s
Block cipher operation modes – CTR
Counter (CTR)
– encrypt – decrypt
– advantages
• efficiency (parallelizable)
• random access (the i-th block can be decrypted independently of the others)
• preprocessing (the values to be XORed with the plaintext can be pre-computed)
• security (at least as secure as the other modes)
• simplicity (does not need the decryption algorithm)
EE
Pi Ci
K
+
(n)
(n) (n)
counter + i
(n)
EE
Ci Pi
K
+
(n)
(n) (n)
counter + i
(n)
P ro to co l p ri m iti ve s
Enveloping
public-key encryption is slow (~1000 times slower than symmetric key encryption)
it is mainly used to encrypt symmetric bulk encryption keys
to co l p ri m iti ve s
generate random symmetric key generate random
symmetric key symmetric-key
cipher (in CBC mode) symmetric-key
cipher (in CBC mode)
plaintext message
public key of the receiver
asymmetric-key cipher
asymmetric-key cipher
digital envelop
bulk encryption key
Message Authentication Codes (MAC)
used to protect the integrity of messages
also called cryptographic checksums
computation of a MAC involves a secret (shared key)
can be based on an encryption function E
Y
1= E
K(X
1)
Y
i= E
K(X
i+ Y
i-1) MAC
K(X) = Y
last or a hash function h
MAC
K(X) = h(X|K)
or both
MAC
K(X) = E
K(h(X))
P ro to co l p ri m iti ve s
HMAC
definition
HMACK(X) = h( (K+ + opad) | h( (K+ + ipad) | X ) ) where
– h is a hash function with input block size b and output size n
– K+ is K padded with 0s on the left to obtain a length of b bits
– ipad is 00110110 repeated b/8 times – opad is 01011100 repeated b/8 times – + is XOR and | is concatenation
design objectives
– to use available hash functions
– easy replacement of the embedded hash function – preserve performance of the original hash function – handle keys in a simple way
to co l p ri m iti ve s
–Digital signatures
similar to MACs but
– unforgeable by the receiver – verifiable by a third party
used for message authentication and non- repudiation (of message origin)
based on public-key cryptography
– signature generation is based on the private key of the sender
– signature verification is based on the public key of the sender
example: RSA based digital signature
– public key: (e, n); private key: (d, n)
– signature generation (input: m; output: )
(m) = md mod n
– signature verification (input: , m; output: yes/no)
e mod n = m?
P ro to co l p ri m iti ve s
“Hash and sign” paradigm
motivation: public/private key operations are slow
approach: hash the message first and apply public/private key operations to the hash only
to co l p ri m iti ve s
hh encenc
private key of sender
message hash signature
hh
message hash
decdec
public key of sender
signature
compare compare generationverification
ElGamal signature scheme
key generation
– generate a large random prime p and select a generator g of Z
p*
– select a random integer 0 < a < p-1 – compute A = g
amod p
– public key: ( p, g, A ) private key: a
signature generation for message m
– select a random secret integer 0 < k < p – 1 such that gcd(k, p – 1) = 1
– compute k
-1mod (p – 1) – compute r = g
kmod p
– compute s = k
-1( h(m) – ar ) mod (p – 1) – signature on m is (s, r)
P ro to co l p ri m iti ve s
ElGamal signature scheme cont’d
signature verification
– obtain the public key (p, g, A) of the signer – verify that 0 < r < p; if not then reject the
signature
– compute v
1= A
rr
smod p – compute v
2= g
h(m)mod p
– accept the signature iff v
1= v
2 proof that signature verification works
s k
-1( h(m) – ar ) (mod p – 1) ks h(m) – ar (mod p – 1)
h(m) ks + ar (mod p – 1)
g
h(m) g
ar+ks (g
a)
r(g
k)
s A
rr
s(mod p) thus, v
1= v
2is required
to co l p ri m iti ve s
How to establish a shared symmetric key?
manually
– pairwise symmetric keys are established manually – inflexible and doesn’t scale
with symmetric-key cryptography
– long-term symmetric keys are established manually between each user and a Key Distribution Center (KDC)
– cryptographic protocols that use these long-term keys are used to setup short-term (session) keys – the KDC must be fully trusted
with asymmetric-key cryptography
– the symmetric key is encrypted with the public key of the intended receiver
– how to obtain an authentic copy of the public key of the receiver?
K e y m an ag e m e n t
y m an ag e m e n t
A, { B, Kab, Ta }Kas
{ A, Kab, Ts }Kbs
A S B
generate Kab
S B
M
(impersonating A and B)
B, { A, Kab, Ts }Kbs { B, Kab, Ts’ }Kas
A, { B, Kab, Ts’ }Kas { A, Kab, Ts’’ }Kbs
...
{ A, Kab, Ts(n) }KbsThe Wide-Mouth-Frog protocol
a vulnerability
The Needham-Schroeder protocol (1978)
Denning and Sacco attack (1981)
– message 3 doesn’t contain anything fresh for B
– an attacker can cryptanalyze an old session key Kab and replay message 3 to B
– the attacker can finish the protocol
– B will think he shares a key Kab with A, but A is not involved at all
K e y m an ag e m e n t
A, B, Na
{ Na, B, Kab, {Kab, A}Kbs }Kas
S A B
generate Kab
{ Kab, A }Kbs { Nb }Kab { Nb -1}Kab
Public-key Needham-Schroeder (1978)
since N
aand N
bare known only to A and B, one may suggest that they can generate a key as f(N
a, N
b)
Lowe’s attack (1995)
A B
{ A, Na }Kb { Na, Nb }Ka
{ Nb }Kb
A B
{ A, Na }Km
{ Na, Nb }Ka
{ Nb }Km
M
{ A, Na }Kb
{ Na, Nb }Ka
{ Nb }Kb
y m an ag e m e n t
generate random number 0 < a < p-1
and calculate A = ga mod p generate random number 0 < a < p-1
and calculate A = ga mod p
generate random number 0 < b < p-1
and calculate B = gb mod p generate random number 0 < b < p-1
and calculate B = gb mod p
calculate
K= Ba mod p = gab mod p calculate
K= Ba mod p = gab mod p calculate
K= Ab mod p = gab mod p calculate
K= Ab mod p = gab mod p
Diffie-Hellman key exchange (1976)
Initially known:
p large prime
g generator of Zp*
A B
Alice Bob
K e y m an ag e m e n t
Man-in-the-middle attack
consider the following protocol
the MiM attack
A B
A, Ka { message }Ka
A, Ka
{ message }Ka
A M B
A, Km
{ message }Km
y m an ag e m e n t
Public-key certificates
a certificate is data structure that contains
– the public key
– name of the owner of the public key – name of the issuer
– date of issuing – expiration date
– possibly other data
– signature of the issuer
issuers are usually trusted third parties called Certification Authorities (CA)
– need not be on-line
certificates are distributed through on-line databases called Certificate Directories
– need not be trusted
K e y m an ag e m e n t
Single CA
every public key is certified by a single CA
each user knows the public key of the CA
each user can verify every certificate
note: the CA must be trusted for issuing correct certificates
problem: doesn’t scale
CA
…
structures
Certificate chains
first certificate can be verified with a known public key
each further certificate can be verified with the public key from the previous certificate
last certificate contains the target key (Bob’s public key)
note: every issuer in the chain must be trusted (CA0, CA1, CA2)
CA1 KCA1
KCA0-1
CA2 KCA2
KCA1-1
Bob KBob
KCA2-1
KCA0
CA structures
CA structures
CA0
CA1 CA2 CA3
CA11 CA12 CA23 CA31 CA32
each user knows the public key of the root CA
0Alice Bob
structures
CA structures cont’d
each user knows the public key of its local CA
CA0
CA1 CA2 CA3
CA11 CA12 CA23 CA31 CA32
Alice Bob
CA structures
CA structures cont’d
each user knows the public key of her root CA
CA1 CA3
CA11 CA12 CA2 CA31 CA32
Alice Bob
structures