The SHA-1 collision project investigates the role that high-performance computing, both software- and hardware-based, can take on for the study of the vulnerabilities of currently used cryptographic hash functions, namely the SHA function family.
In fact, recent results have shown that, based on some sophisticated and computation-intensive techniques, cryptographic hash functions are vulnerable to collision search attacks. This is especially true of MD5, now considered unreliable, while collisions have only been shown for some "weakened" versions of SHA-1. Such results may critically impact today’s security, since most standards, such as DSA and RSA, involve message hashing as the first fundamental step.
Started in 2009 at the University of Naples Federico II, Italy, the project has been supported by the HPC-Europa2 consortium for the development of a high-performance computing application, run at the Barcelona Supercomputing Center (BSC) on a cluster of Cell B.E. multicore/SIMD processors (namely, the Maricel cluster), and by the GARR Italian consortium.
Furthermore, to better understand and demonstrate how the potential vulnerabilities of the SHA-1 algorithm can be exploited, the project has also included the design of a dedicated hardware engine, based on reconfigurable hardware technologies, to speed up the collision search process. As of June 2010, a significant result was already achieved: we were able to find a collision for a 72-round reduced version of the SHA-1 cryptographic hash function, the most advanced result towards the break of the full 80-round SHA-1 at the time of the discover.
We aim in the near future at enlarging the scope of the research, looking at SHA-1’successor, SHA-3, currently being defined. Even more ambitiously, we plan to explore extensively the role of next-generation computing technologies, ranging from processor-based highly parallel systems down to new hardware computational paradigms enabled by emerging trends in information processing technologies and devices. In that respect, the activity presented in this web site will be part of a larger research initiative, addressing the systematic study of present and future technology-enabled cryptanalysis approaches, introduced by the next wave of tera-computing information processing technologies.