可快速定位的视频流媒体大数据存储系统

根据流媒体大数据的特点及其在检索定位、存储容量和速度上的局限性,提出了一种新型的视频流媒体大数据以帧为单位的存储架构;设计了与之相适应的硬件存储阵列系统架构和数据管理系统。重点研究了利用现有IDE硬盘设计形成存储视频流媒体大数据的控制器和硬盘存储阵列系统的结构。利用本文设计的存储阵列结构,一个具有n个硬盘的阵列,在实现n倍扩容的同时可以实现n倍提速;并且可以实现快速定位到帧。系统已经在Xilinx的XUPV5_LX110T开发平台通过验证。

A Video Streaming Media Big Data Storage SystemSupporting Rapid Positioning

Corresponding to the features of streaming media big data and the limitations in retrieval positioning, storage capacity and speed, a novel memory architecture of video streaming media big data stored by frame is proposed in this paper. The relevant hardware storage array system architecture and data management system are designed, with the research emphasis lying on designing video streaming media big data storage controller with existing IDE hard disk and the architecture design of disk storage array system. Adopting the storage array architecture designed in this paper, the storage capacity of an array of n disks expands by a factor of n together with an n-fold improvement in speed, achieving frame fast positioning. The system designed in this study is verified on Xilinx XUPV5_LX110T development platform.