Research On Self-organization And Management Method Of Data In Cloud Storage

With the rapid development of mobile communication and network technology, the amount of information hava explosive growth, and we need more storage space than before. In this trend, the cloud storage service has become a mainstream solution with development of cloud services and the concept of cloud storage services. Users can upload their data to the cloud platform. However, because of user didn’t have the absolute control right about their data in cloud platform and the CPS has the complete logical view, the physical view, and the mapping between two, data secuerity, urse’s privacy and the confidence between user and CPS that hava been focused in recent years.In order to figure out the security issues about usr’s privacy and sensitive data in the traditional cloud storage system. We first address a ASMM(all self-management model), a kind of method of data about self organization and management in cloud storage system, and then conduct research on the key tecniques of this model, including metadata management methods for sensitive and private data, and file autonomos partition startegies, and data distribution algorithms.For the metadata self-management of privacte and sensitive information, ASMM divided metadata management into the client side metadata management (Sub-Metadata Server, Sub-MS) and the cloud side metadata managemen (Core-Metadata Server, CMS). Sub-MS aimed to extract indenpently, maintain, read, write and store the metadata of privacy and sensitive information, which used random redundancy algorithm in maintaining mapping relations, based on the redundancy between files and data blocks, to improve the files security. At the same time, Sub-MS used the request sysnchronism strategy to keep the consistency of mapping relations from data blocks to Data Server(DS) between Sub-MS and CMS, to insure the stability and reliability of the while storage system. To improve the I/O proformance of system and reduce the Sub-MS’s load, we adopted a cache mechanism in the client that ensure the consistency of the cache system through direct synchronization method and LRU algorithm.Aiming at the storage security of the data paritions for private and sensitive files, we first propose autonomos partition algorithms that splits file into many blocks on basis of properties of privacy and sensitive information, and then design the data distribution algorithms that would merge the primitive block set into same number new blocks with Access Structure (AS). The system would store AS into Sub-MS that used to distribute data block.ASMM would ensure the storage security of user’s privacy and sensitive information from files and data blocks two levels with the metadata self-management strategy and data block partition storage mechanism. With multi-user concurrent access, the experiment about proformance of reading and writing for the data sets of different size indicates that ASMM still have good reading and writing performance and data persistence capabilities in managing metadata of privacy and sensitive information by itself and block independent storage compared with HDFS and GFS, and improved the security of storage system.