当前位置:主页 > 科技论文 > 计算机论文 >

面向海量高清视频数据的高性能分布式存储系统

发布时间:2018-03-17 10:43

  本文选题:海量存储 切入点:视频存储 出处:《软件学报》2017年08期  论文类型:期刊论文


【摘要】:通过对视频监控数据的特点和传统存储方案进行分析,提出一种高性能分布式存储系统解决方案.不同于传统的基于文件存储的方式,设计了一种逻辑卷结构,将非结构化的视频流数据以此结构进行组织并直接写入RAW磁盘设备,解决了传统存储方案中随机磁盘读写和磁盘碎片导致存储性能下降的问题.该方案将元数据组织为两级索引结构,分别由状态管理器和存储服务器管理,极大地减少了状态管理器需要管理元数据的数量,消除了性能瓶颈,并提供了精确到秒级的检索精度.此外,该方案灵活的存储服务器分组策略和组内互备关系使得存储系统具备容错能力和线性扩展能力.系统测试结果表明,该方案在成本低廉的PC服务器上实现了单台服务器可同时记录400路1080P视频流,写入速度是本地文件系统的2.5倍.
[Abstract]:By analyzing the characteristics of video surveillance data and the traditional storage scheme, presents a solution of high performance distributed storage system. File storage based on the way different from the traditional design, a logical volume structure, non structured video data stream based on the structure of the organization and write directly to RAW disk devices, to solve the traditional random read and write disk storage scheme and storage disk fragmentation leads to the problem of declining performance. The program will be organized into two level index structure of metadata, respectively by the state management and storage management server, which greatly reduces the number of state manager to manage the metadata, eliminate performance bottlenecks, and to provide accurate retrieval precision seconds class. In addition, the flexible storage server grouping strategies and intragroup preparation makes storage system has the fault-tolerant ability and linear expansion capability. The test results show that the scheme implements a single server on a low cost PC server. It can record 400 1080P video streams at the same time, and the writing speed is 2.5 times that of the local file system.

【作者单位】: 武汉光电国家实验室(华中科技大学);华中科技大学计算机科学与技术学院;
【基金】:国家重点研发计划(2016YFB1000202)~~
【分类号】:TN948.6;TP333

【相似文献】

相关硕士学位论文 前1条

1 陈霄;分布式喷泉码在分布式存储系统中的应用模型与方法研究[D];电子科技大学;2013年



本文编号:1624387

资料下载
论文发表

本文链接:https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/1624387.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户494c8***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com