摘要
随着大数据时代的临近,面对海量数据的统一存储和管理,传统的网络存储方式DAS、NAS、SAN的瓶颈逐步显现出来,DAS只适用于早期内部少量数据存储,SAN和NAS虽然在存储容量和性能上有了较大提升,但受制因素多,对此文中基于Doubl-Cluster双集群架构,结合并行存储等技术来解决上述问题。事实证明该存储方案能有效消除I/O瓶颈,大幅提高聚合带宽和文件查询效率,保证数据共享。
With the approaching of the era of big data,in the face of the unity of the mass data storage and management, the bot t leneck of the t radi t ional network storage way such as DAS, NAS and SAN appears step by step. DAS only applies to the early a small amount of data storage inside,SAN and NAS with large promotion in the storage capacity and performance,but have many limitation factors. This paper based on Doubl - Cluster dubl cluster architecture,combined with the parallel storage and other technology to solve the above problems. The experiments prove that the storage scheme can effectively eliminate the 10 bottlenecks,increase efficiency of the aggregate bandwidth and file query,ensure the data sharing.
出处
《信息技术》
2017年第4期132-135,140,共5页
Information Technology
关键词
网络存储
分布式系统
双集群
元数据
network storage
distributed system
Doubl-Cluster
meta data