COSC 6397 – Information Assurance

1. COSC 6397 – Information Assurance Module M2 – Protocol Specification and Verification University of Houston Rakesh Verma Lecture 1 of M2 (This work is supported in part by NSF) Dr. Verma

2. Contents of M2 • Cryptographic basics • Types of Protocols • Security properties • Taxonomy of Flaws and Attacks • Specification of Protocols • Specification of properties • Protocol analysis Dr. Verma

3. Cryptographic Basics • General principles • Sender, receiver, plaintext, ciphertext, encryption, decryption, etc. • Symmetric key (or secret key) cryptography • Public key (or asymmetric) cryptography • One-way hash algorithms All of these were covered in module M1 ? Dr. Verma

4. Cryptographic Basics (contd.) • Sender – one who sends • Receiver – one who receives • Plaintext – message to be sent, • Notation: P or M • Ciphertext – encoding of P or M, • Notation: C Dr. Verma

5. Cryptographic Basics (Contd.) • Encryption – the process of disguising a message to hide its contents • Notation: E(M) = C • Decryption – the process of decoding C to recover M • Notation: D(C) = M • Basic Identity: D(E(M)) = M Dr. Verma

6. Cryptographic Basics (contd.) • Cryptography – the art and science of keeping messages secure • Cryptographic algorithm – function used for encryption and decryption. • Restricted (secret) or Unrestricted (published) our focus • Unrestricted – based on a key K. EK and DK. The key for encryption and decryption can be different. Dr. Verma

7. Cryptographic Basics (contd.) • Symmetric key cryptography – encryption key can be computed from the decryption key or vice versa. • Special case: the two keys are the same. • Key(s) must be kept secret! • Public key cryptography – encryption key is public the decryption key is not. • Decryption key should be hard to compute from the encryption key! Dr. Verma

8. Cryptographic Basics (contd.) • One-way functions – functions that are easy to compute but hard to invert • Computing f(x), given x, is easy • Computing x, given f(x), is hard This sounds easy, but we have no proof that such functions exist! We will pretend they do. • Trapdoor one-way functions – one-way functions such that • Computing x, given f(x) and some y, is easy Dr. Verma

9. Cryptographic Basics (contd.) • Notation: 1-way for one-way • 1-way hash functions – A hash function that is also a 1-way function. • A good 1-way hash function is also collision-free. • Security of a 1-way hash function is its 1-wayness. Dr. Verma

10. pro·to·colPronunc… (…) n. 1. • The forms of ceremony and etiquette observed by diplomats and heads of state. • A code of correct conduct: safety protocols; academic protocol. 2. The first copy of a treaty or other such document before its ratification. 3. A preliminary draft or record of a transaction. 4. The plan for a course of medical treatment or for a scientific experiment. 5. Computer Science. A standard procedure for regulating data transmission between computers. Dr. Verma

11. Protocols • Protocol – a series of steps involving two or more parties to accomplish a task. • Must be unambiguous • Must be complete in some sense (specified action for lots of possible situations). • Each step is either a computation or a message • Parties may distrust each other Dr. Verma

12. Types of Protocols • Our protocols are cryptographic – use cryptography for preventing eavesdropping, cheating, etc. • Goal of the protocol is beyond secrecy. • Examples: simultaneously sign a contract, convince one another of their identity, etc. • Protocols can be classified in many ways • According to: parties involved, the purpose, the environment, etc. Dr. Verma

13. Classification by Parties • Arbitrated protocols • Adjudicated protocols • Self-enforcing protocols Dr. Verma

14. Arbitrated Protocols • Arbitrated protocols – have an arbitrator, a disinterested third party trusted to complete a protocol. • Easier if parties are face to face. • Over computer networks this results in delay and overheads. • Arbitrator becomes a bottleneck. • Scaling issues. • Arbitrator is vulnerable. Dr. Verma

15. Adjudicated Protocols • Adjudicated protocols – A two stage protocol with: • A nonarbitrated subprotocol • An arbitrated subprotocol executed only in exceptional circumstances – a dispute. • This kind of arbitrator is called adjudicator • Adjudicator only called in to judge fair execution of protocol. Detects cheating rather than preventing. • Good adjudicated protocol – adjudicator should be able to determine cheater’s identity Dr. Verma

16. Self-enforcing Protocols • Self-enforcing protocols – protocol itself guarantees fairness. • No arbitrator or adjudicator – if one party cheats, the others detect the cheating. • Best type of protocol. • Do not exist for every situation. • Exercise: Find a situation for which there are no self-enforcing protocols. Dr. Verma

17. Protocol Classification by Aim • Key-exchange protocols • Authentication protocols • Authentication and Key exchange protocols • Electronic Commerce protocols • … Dr. Verma

18. Key Exchange Protocols • Goal is to distribute keys for secure sessions, channels, communication, etc. • Classical key exchange protocols • TMN • Symmetric Needham-Schroeder • Denning-Sacco • Deployed Protocols • Kerberos IV • SSL/TLS Dr. Verma

19. The TMN Protocol (1990) (Tatebayashi-Matsuzaki-Newman) • Suitable for networks, mobile computing. • Symmetric. Trusted Server S. • Parties don’t have long term keys. • Randomly chosen keys KA , KB , etc. • Standard encryption function E(.), invertible only by server. • Vernam encryption function V(., .) • V(M, V(M, N)) = N Dr. Verma

20. The TMN Protocol • A S : A, S, B, E(KA) • S B : S, B, A • B S : B, S, A, E(KB) • S A : S, A, B, V(KA , KB) • A extracts KB from message 4. • Parties should agree on the session key chosen by B. Dr. Verma

21. An Implementation of TMN • n = p.q, p, q are primes • E(x) = x3 mod n • S knows the 2 prime factors of n • V(x , y) = x exclusive-or y Protocol looks good, but has big flaws! Dr. Verma

22. Authentication Protocols • Authentication protocols – for authentication of parties (principals) • Authentication – assurance of who you are talking to • Examples of specific aims: • To make sure that those obtaining a session key are who they say they are • Make sure that the principal you think has the key does have it. Dr. Verma

23. Authentication Protocols • Passwords or shared keys typically used by system administrators • Authentication can be a byproduct of a key-exchange protocol • Some authentication protocols • Feige-Fiat-Shamir (1987) • Guillou-Quisquater (1988) • Schnorr (1989) Dr. Verma

24. Guillou-Quisquater Protocol • Smart-cards and other applications • Alice wants to prove her identity, bit string J, to Victor • Public information: exponent v, and a number n • (n = p.q, p and q primes) • Private key: B, with JBv = 1 (mod n) Dr. Verma

25. Guillou-Quisquater Protocol • A V : J P wants to prove that this J is hers • A V: T = rv mod n (1 < r < n - 1, r random) • V A: d (0 < d < v – 1, d random) • A V: D = rBd mod n • V computes T’ = DvJd mod n. If T = T’ (mod n), authentication succeeds. Dr. Verma

26. 3 Important Concepts • Security • Privacy • Reliability Dr. Verma

27. Security • Security – the control of information. • Ensures that: • Authorized parties are properly authenticated • Their messages are sent through a network unaltered. • In a secure system the origin, content and intended recipients of a message can be ensured. • Security is not privacy. Dr. Verma

28. Privacy • Privacy – the subject of information can control the information. • Privacy requires security, but security is not sufficient. • Security may preclude privacy! (by assuring that the subjects of information have neither control nor knowledge of the uses of that information) Dr. Verma

29. Reliability • Reliability – provide certainty in the presence of network failures, memory losses and adversaries. • Reliability and security are interdependent. • Reliability is not security. Reliable protocols on unsecure servers provide reliable services to attackers as well as authentic users. • Reliable electronic commerce requires fail-proof transactions. Dr. Verma

30. Security Properties • Authentication – receiver of a message should be able to ascertain its origin. • An intruder should not be able to masquerade as someone else. • Implemented using shared information or ability to prove unique information (PINs and passwords). • Secrecy – confidentiality. If a message is confidential it can be read only by intended recipients. • Eavesdropping is difficult or useless Dr. Verma

31. Security Properties (contd.) • Integrity – receiver of a message can verify that it has not been modified in transit. • Integrity alone is not security. • Availability – a system must be available • availability can be compromised by malicious hackers, network failures or commercial espionage. • Nonrepudiation – a party cannot reasonably claim not to have taken an action. • Example: sender falsely denies sending a message. Dr. Verma

32. Reliability Properties • Atomicity – indivisibility. An atomic transaction either fails completely or succeeds completely. • Consistency – all relevant parties agree on critical facts of the exchange. • Isolation – result of a set of overlapping transactions must be serializable • Durability – a transaction can recover to its last consistent state. Dr. Verma

33. Other Properties Other properties may also be needed. For example, in Electronic Commerce • Certified Delivery • Goods Atomicity • Etc. are also required. Dr. Verma

34. Primary References • Bruce Schneier, Applied Cryptography • Linda Jean Camp, Privacy and Reliability in Electronic Commerce, PhD dissertation, CMU Dr. Verma