New Directions In Cryptography - Papers We Love

I gave a talk about Whitfield Diffie & Martin Hellman’s (1976) “New Directions In Cryptography” paper at Papers We Love Singapore where they announced the Diffie-Hellman Key Exchange.

Video Link: https://engineers.sg/v/601

Transcript

1. New Directions in Cryptography Whitfield Diffie Martin E. Hellman Papers We Love #018 | Vishnu Prem | @burnflare
2. Hello! Vishnu Prem Computer Science @ NUS NUS Hackers Secure message passing got me first interested in technology Used to be HackerspaceSG member in 2009. Gotta fix that This is my first Papers We Love
3. The paper New Directions in Cryptography Published November 1976 on IEEE Transactions of Information Theory NSA claimed they invented it already Whitfield DIffie & Martin Hellman won the Turing Prize in 2015
4. “The development of cheap digital hardware has freed it from the design limitations of mechanical computing and brought the cost of high grade cryptographic devices down…”
5. Conventional Cryptography New Directions in Cryptography
6. Why cryptography? Conventional Cryptography 2 motivations • Privacy - No one else can read my messages except my intended recipient - No one else can inject messages messages to my intended recipients • Authentication - Recipients knows that only sender could have sent this message Most modern standards give both, auth-only standards exists too
7. Alice Bob Eve Secure Channel Insecure Channel
8. One time pad Conventional Cryptography One time pad (OTP) is randomly generated and shared between Alice & Bob OTP needs to be as large as message Unconditionally secure vs. Computationally infeasible
9. Alice Bob Eve Secure Channel Insecure Channel Plaintext: HELLO Key: XMCKL Key: XMCKL Text + Key (mod 26) H E L L O + X M C K L = E Q N V Z Ciphertext: EQNVZ E Q N V Z - X M C K L = H E L L O Text - Key (mod 26)
10. Disadvantages of OTP Conventional Cryptography OTP needs to be as large as the message. Not so practical Instead of Unconditionally secure, let’s use Computationally infeasible instead
11. Symmetric stream cipher Conventional Cryptography Process input in small chunks of text Common stream cipher: Rivest Cipher 4 - RC4 (No longer considered secure) Symmetric means same key for encryption & decryption RC4 HELLO YXTE
12. Symmetric block cipher Conventional Cryptography Acts on large blocks of texts at once Small change in input produces major changes in output This error prorogation is useful: e.g. hash functions Common block cipher: AES-256bit This paper mainly deals with block ciphers AES HELLO YXTE
13. Alice Bob Eve Secure Channel Insecure Channel Key: E8E9EAEBEDE… Plaintext: 014BAF2278A… Ciphertext: 6743C3D151… Ciphertext: 6743C3D151… Plaintext: 014BAF2278A… Key: E8E9EAEBEDE… Ciphertext: 6743C3D151…
14. Threats Conventional Cryptography Cipher only attack • Eve has access to cipher text Known plaintext attack • Eve has access to both plaintext & ciphertext Chosen plaintext attack • Eve has access to cryptographic system
15. Intro to key sizes Conventional Cryptography All ciphers can be brute forced 128 bits -> 128 0s & 1s 2^128 possibilities -> 340 billion billion billion billion If you could test 1 billion keys a second -> 5 * 10^21 years
16. Diffie Hellman (Merkel) Key Exchange New Directions in Cryptography
17. Alice Bob Eve Secure Channel Insecure Channel Key: E8E9EAEBEDE… Key: E8E9EAEBEDE…
18. Alice Bob Eve Insecure Channel
19. Alice Bob Eve Insecure Channel
20. Alice Bob Eve Insecure Channel
21. Alice Bob Eve Insecure Channel + +
22. Alice Bob Eve Insecure Channel + +
23. Alice Bob Eve Insecure Channel + +
24. Alice Bob Eve Insecure Channel + + = =
25. One way function DH Key Exchange A function that is easy one way, but computationally hard the other way DH paper proposed exponents
26. One way function DH Key Exchange Y = a^X mod q q = prime number a = primitive root of q
27. One way function DH Key Exchange Y = 3^X mod 17 Y will distribute within 1-16
28. One way function DH Key Exchange 12 = 3^? mod 17 Discrete logarithm problem
29. One way function DH Key Exchange 3^29 mod 17 12 3^? mod 17 12 Easy! Hard!
30. Alice Bob Eve Insecure Channel 3 mod 17 Secret: 15 Secret: 13 Shared: 12 Shared: 6 12^15 mod 17 = 10 6^13 mod 17 = 10 Alice: 6 Bob: 12 3^13^15 mod 17 = 10 3^15^13 mod 17 = 10 3^13^15 mod 17 = 10 3^15^13 mod 17 = 10 3^15 mod 17 = 6 3^13 mod 17 = 12
31. Public Key Cryptography New Directions in Cryptography aka. the thing you need to do to use github
32. RSA New Directions in Cryptography DH Paper proposed a way public key cryptography could be done 2 years later, RSA was released Asymmetric Ciphers - Prime Numbers • Generate a key pair • One key to encrypt, pair key to decrypt
33. RSA New Directions in Cryptography Asymmetric Cipher generartes two keys: Public & Private key Plaintext + Ciphertext = Ciphertext + Plaintext =
34. Alice Bob Eve Insecure Channel
35. Alice Bob Eve Insecure Channel
36. Alice Bob Eve Insecure Channel Plaintext
37. Alice Bob Eve Insecure Channel Plaintext
38. Alice Bob Eve Insecure Channel Ciphertext
39. Alice Bob Eve Insecure Channel Ciphertext
40. Alice Bob Eve Insecure Channel Ciphertext
41. Alice Bob Eve Insecure Channel Ciphertext
42. Alice Bob Eve Insecure Channel Ciphertext
43. Alice Bob Eve Insecure Channel Ciphertext
44. Alice Bob Eve Insecure Channel Ciphertext
45. Alice Bob Eve Insecure Channel Plaintext
46. RSA New Directions in Cryptography But • Asymmetric ciphers are 1000s slower than symmetric ciphers So we use RSA to just share a symmetric key instead
47. Alice Bob Eve Insecure Channel
48. Alice Bob Eve Insecure Channel Symmetric Key
49. Alice Bob Eve Insecure Channel Plaintext
50. Alice Bob Eve Insecure Channel Ciphertext
51. Alice Bob Eve Insecure Channel Ciphertext
52. Alice Bob Eve Insecure Channel Ciphertext
53. Alice Bob Eve Insecure Channel Ciphertext
54. Alice Bob Eve Insecure Channel Ciphertext
55. Alice Bob Eve Insecure Channel Ciphertext
56. Alice Bob Eve Insecure Channel Ciphertext
57. Alice Bob Eve Insecure Channel Symmetric Key
58. New Directions in Cryptography Whitfield Diffie Martin E. Hellman Papers We Love #018 | Vishnu Prem | @burnflare