Application Study Of Ciphertext Scheduling Algorithm In Cloud Storage

With the development of Internet and wireless sensor network, large amounts of data are being generated, which lead us into a big data era. To process these data, cloud computing emerged. C loud storage is the extension and expansion of cloud computing, it has two demands. First, data is stored in the cloud server by a hosted way. Because the data is out of the control of data owner, it needs to ensure the security of the data by encrypted way. Second, in consideration of disaster recovery performance, storage load balance, hot load balance and the calculation of the efficiency in cloud storage, it needs to make multi-objective optimization of replication node selection according to the feature of the replication.The research is aimed at the two demands of cloud storage and satisfy the Jiangsu Provincial Joint Innovation and Research Funding of Enterprises, Colleges and Institutes—Prospective Cooperative Research Project(BY2013015-23). this paper proposes a replication distribution strategy based on multi-objective decision and a similarity measure method of ciphertext replication based on keywords extracted. In this paper, the main work is as follows:(1)This paper researched the Local Optimum Distribution Strategy(LODS) specific to that optimization goals of the existing replication distribution strategy are too single. LODS makes a series of new definitions, and uses consistent hashing to reduce the scope of nodes in the cloud storage system, then combines AHP. and LODS searches the target storage nodes as candidate nodes in cloud system within a local range by making decisions within a optimized decision radius, then chooses the best target node from candidate nodes by researching multi-objective optimization criterion in cloud storage deeply. The value of the optimized decision radius is relatively stable with the expansion of the scale of cloud storage system and the violent increasement of data. When the decision to take the best decisions radius, performance(storage load balancing, load balancing heat, waiting time) of LODS was higher than the performance of replica distribution policy strategy in HDFS and Amazon S3 system.(2)Existing method of keyword weight qua ntification does not consider the scope attribute of keywords, that can’t make fair quantification of keywords in the document with different length and structure. In response to this shortage, we researched the Correlation Measure of C ipher Text Based on Keyword Re_extract(CMCTBKR). Forming a cloud storage cipher text scheduling algorithm based on the LODS(C-LODS) by using CMCTBKR to measure the correlation, which will join LODS rule layer, between ciphertext replications and storage nodes. The experiment results show that C-LODS can effectively optimize the correlation of balance of replications.