Cracking MD5 hashes by simultaneous usage of Multiple GPUs and CPUs over multiple machines in a network

Cryptographic Hash functions find ubiquitous use in various applications like digital signatures, message authentication codes and other forms of security. Their associated vulnerabilities therefore make them a prevalent target for cyber criminals. General means of defeating hash based authentication involves both discovering and exploiting inherent weakness in cryptographic hash functions or by straight forward brute force attack on the hash Digest. The latter in most cases presents to be an extremely time intensive task. Recent times have however seen usage of GPUs for brute forcing hashes thus significantly accelerating the rate of brute forced hashes. This has further been extended to use simultaneous usage of multiple GPUs over multiple machines. Usage of CPUs for brute forcing have more or less lost importance over the advent of GPU based cracking. This paper presents an efficient method of maximally using computing resources by simultaneously using Graphics Processing Units (GPUs) and Central Processing Units (CPUs) of machines over a network and compares its efficiency against using networked GPUs alone. This also discusses the possibility of currently unseen exploitation of computing power of botnets. Keywords— MD5 Hash cracking, CUDA based cracking, combined usage of CPU and GPU for hash cracking, Botnet