110 likes | 269 Views
Security Considerations for Wireless Sensor Networks. Prabal Dutta dutta.4@osu.edu (614) 975-3658. Overview. Motivation Basic Concepts Confidentiality Authentication & Non-repudiation Integrity Freshness Challenges Key Distribution Strategies Constraints Others Take Away.
E N D
Security Considerations for Wireless Sensor Networks Prabal Dutta dutta.4@osu.edu (614) 975-3658
Overview • Motivation • Basic Concepts • Confidentiality • Authentication & Non-repudiation • Integrity • Freshness • Challenges • Key Distribution Strategies • Constraints • Others • Take Away
Motivation • This Presentation • Introduction to security concepts for 1451.5 survey • In General • Wireless Sensor/Actuator Networks may be collecting sensitive data and controlling important assets that must be protected • Wireless is a broadcast medium so inherently insecure and subject to eavesdropping • Vulnerable to active attacks like message spoofing or message replays • Obscurity IS NOT security • MUST have security for a complete solution
Confidentiality • Allows communicating parties to keep secrets • Protects against eavesdropping • Implemented through encryption using ciphers • Private (“Shared”) Key (Symmetric Cipher) • Fast, small, efficient, but has a key distribution problem • Public Key (Asymmetric Cipher) • Solves key distribution, but computationally very expensive • But encryption is not enough – if the same message encrypts to the same cipher text – then replay attacks are possible (“freshness”)
Authentication & Non-repudiation • Related concepts that allow a message recipient to verify the authenticity of the message source • Allow a party to prove its identity • Protects against forgery or masquerade • If a priori knowledge of peer (and a shared secret) exists, then use of knowledge of shared secrets is sufficient • Otherwise, requires use of a digital signature from a trusted source (like a notary public) • Digital Signatures implemented using a member of the public key cryptography suite (e.g. DSA) - but computationally expensive
Integrity • Ensures that: • Messages are not accidentally corrupted by an imperfect communications channel • Message are not intentionally corrupted by an attacker during transmission • Typically implemented through cryptographic hashes that are similar to a CRC. • Hashes are considerably more complex than a CRC function, and therefore practically impossible to reverse engineer. • Common hashes include MD5 and SHA
Freshness • Ensures that: • Messages are current • Messages are ordered • Messages are not duplicates (replays) of a previously transmitted message • Weak freshness provides partial message ordering but no delay information • Strong freshness provides complete ordering and delay estimation • Implemented through sequence numbers, timestamps, etc.
Key Distribution Strategies • Shared Global Key (pre-deployment) • Frequently used • Compromised node compromises entire network • Shared Pair-wise Keys (pre-deployment) • Requires n(n-1)/2 ~ n^2 keys • Compromised node compromises entire network • New node requires n + 1 new keys n + 1 updates • Ad Hoc Pair-wise Keys (post-deployment) • Computationally expensive (D-H KE: 2.24 J in 5.75 sec on 206MHz SA [Hodjat] -> 1850 reps on CR2450 battery) • Vulnerable to DoS Attacks • Base Station Key Distribution Center (pre & post) • Pair-wise keys between nodes and BS (Kerberos-like) • Ad hoc pair-wise keys between nodes with BS help
Constraints • Limited source of energy • Limited computational resources • MHz • FLOPS • Code size • RAM size • Limited communications bandwidth • Asymmetric ciphers (e.g. RSA) and key exchange protocols (e.g. Diffie-Hellman) are computationally expensive
Other Challenges • Cipher Selection • Must be small, fast, and readily available • Global Clock Synchronization • Required for certain functions like freshness • Authenticated Broadcast [TESLA, TIK] • One node simultaneously broadcasts authenticated message to several other nodes • Also requires clock synchronization • Static vs. Mobile Nodes • Mobility adds another dimension of complexity to key distribution
Take Away • Many constraints & tradeoffs and nothing comes for free • Energy is biggest constraint • Key exchange is a major challenge • No “one-size-fits-all” solution • Active area of academic research • Berkeley, UCLA, Network Associates, others