Design Of Cloud Storage Backup System Based On P2P And Implementation Of Log Recovery Module

As the development of computational world, cloud computing has become a popularcomputing model to support processing massive data using clusters of commodity computers.Cloud storage is a new concept developed from the concept of cloud computing, which isattracting the attention of academia and industry. Enterprises backup data in the remote datacenter using cloud storage, in the case of a major disaster, and enterprises quickly realize theremote disaster recovery through backup data in the cloud storage system to avoid the loss ofdata, which is not only easy to achieve massive information storage and management, but alsoreduce the cost of enterprises.In the current cloud storage system, the center server plays a crucial role, includingactual storage and access requests from users and so on. In the centralized architecture, thereis not only single point of failure, but also the center server is able to become the performancebottleneck of cloud storage system. If the system collapses, the center server will most likelylead to paralysis of the entire cloud storage service. Even though there is the backup strategyin the cloud storage system, this problem is not resolved totally. In addition, because of poorscalability of the system, with the increase of load balance, the pressure of system willbecome bigger and bigger, but the features of P2P, which are decentralization, robustness,load balance, and scalability, provide a solution for this. By using integrate cloud storage andP2P technology, load is distributed to multiple service management nodes, and we couldremove the fragile center server of the cloud storage, which largely improves the servicequality and performance of cloud storage, so as to enhance the service reliability andavailability of the system.In order to solve this problem, this paper presents the following work:1. Integrate cloud storage and P2P, this paper presents cloud storage backup systembased on P2P, and states the system structure, network topology structure, and the overallframe structure. Chord algorithm is used to coordinate multiple service management nodes,and distribute requests from users to multiple data chunk servers. A service management node,multiple data chunk servers and backup servers constitute a point storage cluster, whichcompletes user data storage and management, and solves the bottleneck of the center node.2. Design file storage and access plan, define the key data structure of file metadatainformation table, file index table, the backup file information table and so on. Analyze theavailability and consistencies of replica, and the appropriate replica management strategy isapplied in this system, and develop a backup and recovery plan, which makes cloud storage backup system based on P2P has a good disaster recovery performance.3. Research and analyze the conventional log structure of the current database system,and put forward a kind of log data structure suitable to the cloud storage backup system basedon P2P, and the log is used to achieve the backup and recovery, so as to ensure the reliabilityand availability of the system. Design the log management engine, and analyze the logrecovery module, including data structure, cache, data types of resource management, as wellas the related manipulation functions of the log, integrate the basic principle and realize thealgorithm of the log recovery, finally, the implementation of log recovery module is given.4. Build the simulation platform, design the process, and analyze the results. Simulationexperiments include the test of backup and recovery, how the number of storage node andbackup replica impact the availability of the system, how the number of backup replicaimpacts the performance of the system, and the success rate of the file recovery. Theexperimental results show that the cloud storage backup system based on P2P has highperformance and usability, and could provide higher quality of cloud storage service forclients.