Analysis of ARX round functions in secure hash functions | Posted on:2012-12-12 | Degree:D.Sc | Type:Dissertation | University:The George Washington University | Candidate:McKay, Kerry A | Full Text:PDF | GTID:1458390008499274 | Subject:Computer Science | Abstract/Summary: | | A new design paradigm for symmetric-key design primitives, such as hash functions and block ciphers, is the Addition-Rotation-XOR (ARX) paradigm. ARX functions rely on the combination of addition modulo 2 n, word rotation and exclusive-or to increase the difficulty of applying traditional linearity-based attacks. This work provides contributions in the analysis of ARX functions.;This dissertation introduces a new analytic technique, pseudo-linear cryptanalysis, which takes advantage of linear properties of ARX-functions over the groups Zn2 and Z2n. This is in contrast to traditional linear analysis, which has largely focused on linearity over Z2. Pseudo-linear cryptanalysis can be used on any ARX-based symmetric primitive, and is particularly useful for block ciphers and iterative hash functions containing round functions. The dissertation also presents a variant that can be used for differential attacks, and extends the branch number diffusion metric for ARX ciphers that use large words.;Secure hash functions are among those primitives that may be built on ARX-functions. The National Institute of Standards and Technology is currently in the process of selecting the next US standard secure hash algorithm, SHA-3, which will be used in everyday applications such as secure online sessions and password-based authentication. Two of the five finalists in the SHA-3 competition are based on ARX functions. This dissertation applies pseudo-linear cryptanalysis, truncated differentials, and new ideas for computing branch numbers to SHA-3 finalist Skein. It also presents improved attacks on second-round SHA-3 candidate CubeHash as well as structural analysis of its symmetry classes. | Keywords/Search Tags: | ARX, Hash, Functions, SHA-3 | | Related items |
| |
|