Security Analysis Of Hash Functions

A cryptographic hash function is a one-way mathematical function that takes a message of arbitrary length as input and produces an output of fixed (smaller) length. Hash functions are the essential building blocks in modern information security systems. The overall security of such systems is crucially dependent on these mathematical functions, which makes the analysis of hash functions critically important. However, the security argument for the majority of hash functions in use is only a heuristic one and therefore their respective security evaluation continually remains an open question. In this thesis, we provide cryptanalytic results for two standardized cryptographic hash functions.The first part of this thesis is focused on the analysis of the hash function HAS-160, the Korean industry standard. By using the differential meet-in-the-middle and the initial structure techniques, we propose improved preimage attacks against step-reduced HAS-160. Our work generates a preimage for70-step HAS-160faster than exhaustive search and also improves the complexities of previous attacks on65,67and68steps. Finally, we provide a security bound for preimage attacks on full HAS-160with the accelerated brute-force search. As far as we know, our result on70steps of HAS-160is the best preimage attack in terms of attacked steps.The second part of this thesis is focused on the fault attack against NIST (National Institute for Standards and Technology, US) hash standard SHA-256. We propose an algebraic fault attack on SHA-256compression function under the word-oriented random fault model. By injecting about65faults, the chaining value and the input message block can be recovered with about200seconds on average. To the best of our knowledge, this is the first algebraic fault attack on hash functions. Moreover, we extend this attack to HMAC-SHA-256and launch an almost universal forgery attack on it. Our algebraic fault attack is generic, automatic and has the potential to work well for other ARX-based primitives.