针对云存储空间数据高冗余问题,提出一种基于擦除码的高效云存储方案。设计一种云存储系统模型,使用擦除码,将编码后的原始数据文件分散存储于云存储系统各节点上,由接收者解码完成数据恢复,保证数据文件高可用性并优化存储冗余度,降低存储空间消耗,利用Chord环提升分布式容灾存储数据块的检索效率。搭建Hadoop云存储平台进行测试,测试结果表明,该方案在保持高可用性的同时,提升了存储空间利用率。
To solve some high-redundancy issues of cloud storage space, a cost-effective cloud storage data redundancy scneme based on erasure coding was proposed. A cloud storage system model was designed, the original data files were stored in the different nodes of the cloud storage system after being coded by erasure coding. The operation of decoding was done by receivers to complete data recovery, which not only ensured the high availability of data file, but optimized the storage redundancy through coding redundancy, thereby reducing the consumption of storage space. Meanwhile, to improve the retrieval efficiency of the data blocks of the distributed disaster tolerance storage, Chord ring was designed in the proposed storage scheme. The storage scheme was implemented and evaluated on Hadoop platform. Results show that it improves the utilization of storage space as well as the availability.