Eraser-based Encoding And Error Correction Algorithms In Distributed Storage Systems

Distributed storage systems(DSS)use replication and erasure code as two redundant strategies to ensure the reliability of the system.The utilization rate of the copy space resource is low,while the traditional erasure correction code is based on the high-advanced field operation,and the coding and decoding complexity is high.The ZigZag Decoding(ZD)decoding process is based on the back-examination XOR operation in the binary field,and the decoding complexity is low.ZigZag decoding has a combination property(CP),that is,k original data blocks are encoded into n data blocks,and as long as any k data block in the n data blocks are taken,the original file can be successfully restored.CP-BZD code is a storage code that has both CP and ZD properties and is widely used in distributed storage and wireless communications.So the research based on CP-BZD code such erasure code coding and error correction algorithm is of great significance.This paper finds that when a single CP-BZD code node stores multiple files,once the node fails,all the data blocks in the k surviving nodes need to be downloaded to be successfully repaired,and the node repair bandwidth is larger.For the problem that the CP-BZD code repair bandwidth is large,three new coding schemes are designed in this paper without increasing the storage overhead.In this paper,the coding and repair process of each scheme are described separately.It is found that the designed scheme not only increases the decoding complexity,but also effectively reduces the repair bandwidth of the failed nodes.Comparison of data results shows that the node repair bandwidths of the three coding schemes designed in this paper are all smaller than the CP-BZD code.With the increase of(n,k)value,the average repair efficiency of the three coding schemes designed in this paper is getting higher and higher.In this paper,by studying the decoding process of CP-BZD code,it is found that its decoding mechanism has wrong diffusion and reduces the reliability of data.At present,there are two error-correction design directions for distributed storage systems: cryptography and channel coding,which require an increase in overhead bits to effectively perform error correction.Therefore,this paper proposes an algorithm based on CP-BZD code with error correction capability.This algorithm does not need to increase other overhead bits,and obtains the error correction ability,and the module of the algorithm can be chosen freely.The main construction process of this algorithm is as follows: firstly,the check equation is constructed by using the coding structure of CP-BZD code,then the Tanner graph is constructed,then the hard decision algorithm is used for iterative decision,and finally the ZigZag decoding recovers the initial information.The result data shows that compared with the ZigZag decoding,this algorithm has better error correction capability in distributed storage systems.In this paper,roadside units are used for distributed data storage,and a vehicle communication network collaborative content delivery framework is proposed to solve the vehicular Ad Hoc Network(VANET),because the network coverage time is short or the vehicle speed is too fast,the roadside unit The transmitted data cannot be successfully received by the vehicle at one time.Then the error correction algorithm proposed in this paper is applied to this VANET framework to solve the problem that the vehicle receiving data is incorrect.The result data shows that combining the proposed framework and error correction algorithm into VANET can increase the probability of successful data reception by the vehicle to some extent.