Parallel Optimization And Realization Of AES Based On Multi-core

Along with the rapid development of computer technologies and communication technologies, the information security will achieve more and more attention. The traditional data encryption standard-DES is replaced by a new encryption standard-AES for its low strength of encryption and vulnerability. Multi-core processors are increasingly popular, then developing parallel software by taking full advantage of the high-performance of multi-core and replacing the serial software gradually will be the development trend of software industry. At the same time, the development of multi-core processors can’t be separated from the application of parallel software. Therefore, the parallel design and realization of the AES algorithm based on multi-core have great significance for strengthening the computer network security and improving the efficiency of network communications.Firstly, some key technologies were summarized in this paper which will be used in the process of parallel optimization of AES, including multi-core, multithread, parallel computing, and some of the mathematical knowledge. Based on these technologies, the overall structure of the AES algorithm was described in detail. Then it carried out a detailed analysis of the key schedule and round transformation, at the same time, it discussed the decryption process of the algorithm. Starting from the analysis of the core component of AES-the round transformation, two methods that adopt data flow decomposition and data decomposition to optimize AES are presented, the data decomposition contains two methods:the decomposition on the input plaintext and the decomposition on each constitute transformation. It described the decomposition process of these two methods in detail and also gave the feasible analysis.According to the character of AES algorithm, three specific decomposition methods were used in this paper. We look forward to, each constitute transformation can execute without having to wait until the above constitute transformation finished by adopting data flow decomposition on constitute transformation, each plaintext data block’s encryption and decryption achieve at the same time by the decomposition on input plaintext, and each constitute transformation can execute in parallel themselves by the decomposition on each constitute transformation. In this paper, the three decomposition methods were implemented and verified by experiments, experimental results showed that, the method of the data flow decomposition on each constitute transformation of AES algorithm can fairly improve the encryption and decryption performance, the method of the data decomposition on input plaintext can improve the encryption and decryption performance significantly, the proportion of the performance improvement increased with the expansion of data size, the method of the data decomposition on each constitute transformation of AES algorithm can’t improve encryption and decryption performance. In a word, the performance improvement of the parallel optimization to AES algorithm mainly reached the expected goals.