The Research Of Remote File Synchronization Model In Cloud Storage Environment

With the advent of mobile computing, the development in computer hardwaretechnology and the growing needs of the user application, more and more individuals havemultiple intelligent computing devices. People’s daily work and life are dependent on themassive information and data in these computing devices, so sharing and managing the filesin multiple devices safely and efficiently is very important.Remote file synchronization service can automatically maintain file consistencybetween multiple devices. Cloud storage technology and NoSQL technology provides asecure, scalable way of backup storage for user’s files. With the demand of file sharing andmanagement between multiple devices of an individual in the era of mass information, thispaper proposes a remote file synchronization model in cloud storage environment-SyncCSmodel. This paper focuses on the system architecture, synchronization protocol, as well asback-end storage environment, the main completed works include:1. By combining the remote file synchronization and cloud storage technology, andtaking into account the deficiency of traditional peer-to-peer file synchronizationmodel, we proposed a C/S pattern remote file synchronization model in cloudstorage environment-SyncCS, and built a three-layer architecture.2. We designed a two-stage file synchronization protocol, which divides the filesynchronization process into two stages: non-real time two-way synchronizationstage and real-time one-way synchronization stage, to meet different user demandsof two-way synchronization and real-time synchronization. Afterwards, weadequately analyzed and discussed all of the conflict situations that may occur in afile synchronization process, and proposed the corresponding conflict resolution.3. With the current hot technology of cloud storage and NoSQL database, we studiedand structured the back-end cloud storage environment of the SyncCS model, andput forward the adaptable metadata storage and file storage solutions respectivelyaccording to the actual characteristics of the data in the SyncCS model.4. Contrapose the redundant synchronization operations which are caused byredundant log, we put forward an optimized solution of client log preprocessing.Contrapose the metadata server inefficiency which is caused by large numbers ofconcurrent accesses, we proposed a metadata storage optimization solution basedon Hbase. We verified the efficiency of the two optimized solutions byexperiments, thus improved the performance of the SyncCS model from the clientside and server side respectively.