Research On The Optimization Method Of Cloud Storage Security Merging Reputation Evaluation

Cloud computing is built on virtualization, it provides a virtual computer withsuper computing resources which integrated distributed computing resources such asinfrastructure, platforms, software and other services via the Internet. Cloud storage asan extension of the concept of cloud computing in recent years has become a hotresearch in academia and industry. Cloud storage provides an effective solution to theonline application, which needs to storage mass data and the storage space often floatup and down. It not only saves any costs, but also provides a larger data center forremote backup and online storage access. Nevertheless, there are also facing manysecurity issues. When users put their sensitive data to the cloud, they lose the right ofdirect control about themselves data. What security threats for the cloud data that mayoccur cannot be perceived, it makes the data’s owner cannot trust the cloud service.In this paper, we concentrate on the mutual trust of the cloud storage services andthe users. Based on the analysis of existing cloud storage security technology, weconverge the reputation mechanism to provide reliable protection for cloud storageservice. And increase trusted third party to ensure the safety monitoring for theparticipants. Finally, while ensuring the cloud data high availability, we optimized datastorage redundancy. The main research work includes the following aspects:(1) The existing cloud storage technologies were analyzed, we comprehensiveand systematic exposit the relate research progress. The main research methods and therelated technologies were discussed, and the principal problems of the current cloudstorage were concluded.(2) The reputation mechanism for cloud storage service was come up. On onehand, since the cloud service has massive data, strong dynamic features centralizedmanagement is so complex. We converge the autonomic computing ideas make thereputation mechanism for cloud computing with the autonomy. On the other hand, due to the existence of the deviant evaluation, a reputation revised system was put forward.This method can effectively identify and filter unfair rating, enhance the accuracy ofreputation calculation.(3) The cloud storage model of a fusion of the reputation mechanism was putforward. It makes the user could obtain reliable and high quality storage service fromthe providers, while it establishes direct mutual trust between the user and serviceprovider. The proposed model support dynamic operation on the data block level(including modify, insert, delete, and append).It also supports the authorized users toaccess the outsourcing latest version data and allow the owner to grant or revoke thedata access.(4) To maintain high availability of cloud data among distributed elastic nodesand reduce the storage redundancy, we proposed a new dependable cloud storagescheme based on erasure codes. Through the coding redundancy to ensure highavailability and optimize the data storage redundancy. As for the retrieval of the datablock on distributed nodes, we use the chord loop to improve the efficiency of search.