The Study And Implementation Of The Parallelization Of The Improved AES Encryption Algorithm For Cloud Computing

The cloud is full of important and sensitive data,the disclosure,destruction or theft of such data has varying degrees of consequence,then ensure the confidentiality and integrity of data in the cloud computing environment is a hot spot in the current research.Using encryption technology can turn plaintext into meaningless ciphertext,the cracking of ciphertext is a time-consuming and difficult project,and for some encryption algorithms,there is no attack method to break them,so encryption technology can improve the security of cloud data.However,the computational complexity of encryption algorithm is generally high,and the time and calculation cost when decrypting big data are very high.Therefore,it is of great significance and value to study the encryption algorithm which is fast to decrypt,high security and suitable for cloud computing environment.The main research content of this paper is how to make use of the advantage of rich computing resources and support virtualization and parallelization of cloud computing,combining encryption to enhance the security of cloud computing.It is found that the AES cryptography algorithm in the symmetric cryptography system is more suitable for decrypting big data than other cryptographic algorithms,and it easy to parallelize.Therefore,AES cryptography algorithm is chosen as the data encryption scheme for cloud computing environment.Considering that the S-box in the traditional AES cryptography has the disadvantage of shorter iterative period and weak nonlinearity and strong correlation between wheel key and seed key.Based on the current situation of domestic and foreign research on AES improvement algorithm,this paper optimized the S-box and improved the key stream generation process with the advantages of stream cipher.By analyzing the experimental results,it is found that when the same plaintext is encrypted multiple times,the key stream sequences and the ciphertexts are different for each encryption.Thus indicating that the improved AES cryptography algorithm achieved the effect of one-time encryption using a seemingly random fixed key.The improved AES encryption algorithm is more secure than the traditional AES encryption algorithm for it has both the security of block cipher and stream cipher system,and the function of generating key flow is a one-way function which is very difficult and even impossible to reverse crack.In order to meet the needs of big data encryption in cloud computing,two parallelization schemes to improve AES encryption algorithm are designed.The first set of scheme is based on multi-core computer platform,creating multiple computing processes and encrypting block data in parallel.The second is to use MapReduce parallel mode,build multiple Map modules for parallel encryption.Ideally,the acceleration ratio of the first scheme is close to the number of computers,and the second is approximately one-third of the number of processors.Through the experimental results of encryption time,parallelization acceleration ratio and CPU utilization,the acceleration ratio based on 4-core computer reached 3.49,and the other one based on 8 processor hadoop cluster reached 2.41,both of which are very close to the ideal acceleration ratio.It is concluded that the parallel improvement AES encryption algorithm has the advantages of fast decryption speed,high security and high resource utilization,and is more suitable for data encryption in a cloud computing environment.