Research On Optimization Algorithm Of Gateway Dump Based On Hybird Cloud Storage

With the rapid development of information technology,and the data volume rendering of explosive growth, file number is increasing and The file is a lot of more phyletic. Now how to deal with the enterprise the peak of storage and preservation of confidential data has become an urgent problem. Hybrid cloud storage is a mixture of public and private cloud, it can solve the problem of the lack of enterprise storage capacity, and provides a solution for the preservation of personal data and common data. Serving as a link between the public and private cloud data transmission and exchange, the gateway of hybrid cloud storage is the key part of hybrid cloud storage and its performance directly affects the performance of hybrid cloud storage. Therefore, studying and optimizing the archived algorithm of hybrid cloud storage gateway have the theoretical and practical significance.This paper adopt a new approach to dump the conversion algorithm of hybrid cloud storage gateway. According to the attribute characteristics of the files in hybrid cloud storage and user’s network relationship,this paper respectively designed similar TBF-K algorithm based on classification decision tree and PRE-SN algorithm based on user network relationship. It provides a forecast for file access behavior and a basis for the cache in the cloud storage gateway archived. The leaf node of the decision tree is a collection of documents with similar properties.Similar TBF-K algorithm based on classification decision tree is mainly to establish the file storage location prediction model which can provides decision basis for the unclassified documents archived based on metadata file. It first uses ID3 algorithm to get the file attribute information gain and to establish the optimal file attributes classification decision tree. Secondly, it calculate the similarity of the new unclassified documents and files in the collection, and it take the highest similarity K file categories as the class attribute of unclassified document.PRE-SN algorithm based on user’s network predicts the hot file through the analysis of the user of a hybrid cloud storage relation. It uses relative centrality to represents the user’s network relationship. At the time of transfer, the higher the degree of user’s relative centrality, as much as possible dump the user’s file to local. On the contrary, if user’s relative centrality degree is not high, it is all or most that transfer the files to public cloud storage.This topic for shooting and time performance of the algorithm takes a complete test through constructing test environment. The results show that the classification decision tree algorithm based on the file metadata and archived algorithm based on the user’s network that all well balance the relationship between storage capacity and storage efficiency.It has made the reading and writing performance of a hybrid cloud storage system greatly enhanced. Therefore archived algorithm designed in this paper, which can effectively improve the processing performance and user experience of a hybrid cloud storage.