Research On The Method Of Cloud Data Accurate Restoration

Cloud storage is a new technology of network storage.Cloud users can storage and access data in the cloud at any time and anytime by connecting to the cloud through networking devices.Cloud storage develop rapidly with characteristics of high efficiency,convenience and low cost.However,cloud users send data to the cloud for storage and maintenance,which make them lose control of data.The following security issues may appear.Cloud service have the problem of node faults and service interruption.More likely to cause data loss,which will cause cloud storage data to be very risky.At present,the general method to solve these problem is to use a multi-cloud environment and redundancy technology to improve the availability of data.Previous researches have achieved to recovery data fast and accurately with less data redundancy and less repair cost,and improve the availability of data.But distributed storage the original data can not protect the data privacy.And the research can only repair a block of data,but not consider the redundant data block repair.In view of the above shortcomings,this paper proposes a new cloud data accurate repair scheme---Z 'code.Firstly,the scheme blocks and hidden-encodes data and generates redundant blocks;Then the data blocks are stored into a plurality of cloud service providers.This scheme solves data confidentiality and data availability,and can use the data that is not lost to repair precisely two pieces of lost data,including data blocks and redundant blocks.In addition,the scheme introduces a trusted third party,to achieve a MapReduce algorithm based on Hadoop distributed system.The MapReduce algorithm parallelize the matrix calculation to hide and encode data for multiple users at the same time,which solves the data confidentiality and reduces the burden of users.The experimental results show that the time of repairing a block of hidden data is slightly different from the existing research.However,the time to repair a piece of redundant blocks is less than the time they repair data block;Moreover,the computational overhead of data hiding and anti hidden in Hadoop parallel working environment is muchimproved than that of the serial algorithm.And the burden of users are lighten.