Research For Rapid Recovery Of RAID-6 Code Extension Technology

The data explosion in the era of big data has brought huge challenge to data storage.On the one hand,the failure probability of new storage media as well as traditional disk surging with the memory size,leads to a normalized level of equipment failure.But the existing codes which can not meet reliability requirements create a demand for seeking more reliable codes.on the other hand,storage System performance lagged far behind the rising storage capacities,and reconstruction time is too long;in this context,this paper aims at the three existing fault tolerance encoding.for example: the poor rebuilding performance,long waiting time,inefficient redundant use,studies the problems of storage reliability from the angle of encoding,combines the array code with partial coding in order to meet the actual needs,and proposes a new extensive coding idea which extended based on two common codings.The main work and innovation are as follows:First,it is found in the study that the existing erasure encoding system only provides extended reliability and does not improve performance,it also did not concern the problems of failure probability such as: the third redundant disk of redundancy check in RTP coding only funtion as the three disks fail at the same time but this probability of occurance is less than 1%;what is more,the existing erasure code does not take into account the reliability of the system when it is in a degraded state.To solve the problems above,this paper proposes a coding for fast recovery RAID-6 expansion ideas.This idea concerns a high probability of a single disk failure situation and accelerates its recovery rate.Method of adding redundancy to the binary addition of redundant group as the basic unit,further reducing the required recovery while reading data blocks meet when the maximum number of data blocks to reuse a single equipment failure,thereby reducing the overhead of data block transfer,improve the recovery efficiency,reduce the time of the downgrade,a substantial increase in the recovery rate while increasing the reliability of the original RAID-6 system to some extent.Secondly,this paper implement the encoding idea above to practice and extend to RDP plus coding,FR2 E coding based respectively on RDP coding and EVENODD encoding.The spreading code in the paper has advantage over the original three fault-tolerant coding by rigorous theoretical analysis.Analysis of the RDP plus spreading code obtained in fast recovery performance improvements up to 18.5%;while the EVENODD FR2 E encoded by a unique coding adjustment factor into the extended redundancy to make recovery performance further improved to reach 21.6%.In addition,this paper also brings out a new metric: Unit redundancy of the performance improving factor Q,namely: the erasure codes performance and storage overhead ratio,used to measure the cost-effective method of adding redundancy.Finally,the paper simulate and implement the process of encoding and decoding using the two spreading codes.The propose extension of coding outperforms the original three fault-tolerant codings by effectively improving the recovery performance compared and reducing the coding overhead.The experiment results further demonstrate the superiority of this method.