CAJ | 학술논문

　　文章根据云存储的结构特点，将数据存储区域划分本地副本区、本地编码区、远端编码区。本地编码区和远端编码区采用确定型网络编码来实现数据的冗余备份。将解码过程的所有系数矩阵进行三角分解并保存为查找表，采用查表方法辅助方程组求解，减少计算量。实验结果表明，当单个存储结点的可用性在0.7至0.99之间、数据可用性目标为99.99999%时，采取数据分片数为8的方案，该方法的平均每字符运算次数低于3次，比副本冗余方法节省50%以上的存储空间.分区编码冗余方法解码速度快，能够显著提高云存储系统的可用性和存储效率.
　　문장근거운존저적결구특점，장수거존저구역화분본지부본구、본지편마구、원단편마구。본지편마구화원단편마구채용학정형망락편마래실현수거적용여비빈。장해마과정적소유계수구진진행삼각분해병보존위사조표，채용사표방법보조방정조구해，감소계산량。실험결과표명，당단개존저결점적가용성재0.7지0.99지간、수거가용성목표위99.99999%시，채취수거분편수위8적방안，해방법적평균매자부운산차수저우3차，비부본용여방법절성50%이상적존저공간.분구편마용여방법해마속도쾌，능구현저제고운존저계통적가용성화존저효솔.
According to the structural characteristics of cloud storage, the data storage area is divided into three parts: local copy zone, local coding zone and remote coding zone, the latter two using a deterministic networkcoding to achieve the purpose of data redundancy. Coefficient matrix of the decoding process triangular decomposition and save the results as a search table that can assist equation solving, thereby reduce the amount of calculation. The experiment results show when the availability of a single storage node in the range of 0.7 to 0.99, data availability target of 99.99999%, taking the data slice number 8, the average operations per character less than 3 times, the method can save more than 50% of storage space than copy redundancy method. The partition coding redundancy method has high decoding speed, improves the availability and storage effciency of cloud storage system signifcantly.