Distributed Multi-core CPU And GPU WPA/WPA2-PSK High-speed Cracking

With the development of the information society, wireless hotspot becomes widespread in our life. According to the forecast of the World’s Wi-Fi Alliance, the total number of world’s home WiFi networks will be close to 800 million in 2016. Most of the networks use WPA/WPA2-PSK algorithm to encrypt WiFi, which makes it more valuable to make research on cracking WPA password to improve the safety of WiFi. In order to overcome the limitations of the existing WPA/WPA2-PSK decoding method based on single core CPU or multi-core CPU, this paper proposes a new parallel method for the distributed multi-core CPU and GPU. In the hybrid system,the PSK list is distributed to every computer using reasonable distributed technology. Computing tasks are performed in parallel by multi-core CPU and GPU. Due to the high performance computing power of the dense parallel tasks, GPU greatly improves the cracking speed. The experimental results show that the hybrid system can improve the speed of the single core CPU by more than two orders of magnitude, and present a more significant advantage and distributed performance. This thesis also attempts to call cloud computing resources through the API interface to provide greater computing power for the hybrid system. In this thesis we propose an improved Amdahl rule to analyze the cracking speed limit. In the experiment, through the analysis of the influence graphics parameters, we can clearly recognize the importance of choosing the appropriate graphics card.