On the Performance of Cracking Hash Function SHA-1 Using Cloud and GPU Computing

This study cracks code words by matching them with target hash values using cloud distributed and GPU parallel computing technologies. For the cloud computing system, this paper implements the matching algorithm using a Hadoop distributed system consisting of four physical computers, one used for the master node and the other three used for worker nodes. Also, MapReduce is used in nodes to handle large amounts of operations. For the GPU parallel computing system, this paper utilizes thousands of GPU threads to operate in parallel on the data allocated by the CPU. The experimental results of the cloud distributed system show that a cloud with more than 330 computing nodes can crack hash values of code words with length of 7 characters within 1 min. Furthermore, based on the results of the GPU parallel computing system, each thread can execute about 1100 hash value matching operations per second; hence the GTX650 Ti, which owns 8192 threads, offers cracking capability comparable to a cloud system with 120 worker nodes.