作为新兴的互联网多媒体服务,视频点播极大的满足了人们对视频信息的需求,在很多领域都得到了广泛应用。
As an emerging internet multi-media service, the VOD (video on demand) greatly satisfies people’s demand on video information, so it has been widely used in many fields.
视频点播系统对存储具有较高的要求,因为存储系统是所有视频资源和多媒体数据的载体,存储系统的性能很大程度上决定了视频点播系统的性能。云存储技术的出现为构建视频点播存储系统提供了新思路,开源云存储软件HDFS的特点完全满足视频点播系统的需要,并且能够为用户节约大量存储设备采购资金。
The VOD system has high requirements on the storage, because the storage system is the carrier of all the video resources and multi-media data, and storage system’s performance largely decides the performance of VOD system. The technology of cloud storage presents a new approach to build the VOD storage system. HDFS,a cloud storage software of Open Source, has features completely meeting the requirements of VOD system, and can save substantial storage device cost for users.
为降低视频点播系统的建设成本,构建高效、低价的视频资源存储系统,本文将云存储技术引入到视频点播领域,应用开源云计算软件Hadoop的分布式文件系统组件HDFS建立文件集群,存储视频点播系统的海量视频数据。
In order to reduce VOD system cost and build a high-efficient and low-cost video resource storage system, the paper introduces the technology of cloud storage into the VOD field, using HDFS, the distributed file system component of Open Source cloud computing software Hadoop, to establish a file cluster to store the huge amount of video data of VOD system.
论文从分析视频点播系统的存储需求入手,阐述了HDFS组件适用于视频存储系统的技术优势,提出了一种基于HDFS的视频点播系统的设计方案,该方案结合了流媒体技术和云存储技术的优势,构建了基于Web的视频点播系统。
Starting from analyzing the storage requirement of VOD system, the paper explains the technological advantages of HDFS component applicable to video storage system, and put forwards a HDFS-based VOD system design plan. The plan builds a web-based VOD system combining the advantages of streaming media and cloud storage.
该系统采用B/S结构,由客户端、Web服务器、流媒体服务器和HDFS集群组成,客户端通过浏览器内嵌流媒体播放器为用户提供视频点播界面和播放视频节目;Web服务器响应用户请求提供视频节目信息;流媒体服务器响应用户视频播放请求,对视频文件进行技术处理生成流媒体数据后发送到客户端;HDFS集群是由廉价硬件构成的分布式云存储系统,负责存储海量视频文件。
The system adopts a B/S structure composed by the client side, web server, streaming media server and HDFS cluster. The client side offers users the VOD interface and plays videos through the streaming media player embedded in the browser; the web server offers video information responding to user requests; the streaming media server makes technical processing to video files responding to the users’ video play requests and sends the streaming media data generated to the client side; the HDFS cluster is a distributed cloud storage system composed by cheap hardware, which is responsible for storing an immense amount of video files.
本文所设计的系统在可靠性、安全性、可扩展性、经济性上均具有一定的优势:
The system designed in the paper has certain advantages in terms of reliability, security, expandability and economical efficiency.
可靠性:系统结合流媒体技术和HDFS云存储技术进行设计,流媒体技术技术成熟,而HDFS技术也通过了上千个结点的运行测试,并且成功地应用于多个知名大型系统中,两种成熟可靠的技术是本系统可靠性的强大保证。视频文件在HDFS集群中以多副本的形式进行保存,同时还可以根据视频文件的重要程度,增加副本的份数,进一步提高了文件存储的可靠性。
Reliability: The system makes design combining streaming media technology and HDFS cloud storage technology. The streaming media technology is mature, and the HDFS technology has passed the running test of thousands of nodes and has been successfully applied to many famous large systems, so the two mature and reliable technologies form a strong assurance for system reliability. The video files are mostly stored in the HDFS cluster in the form of copy, and the number of copy can be increased according to the importance of video files to further increase file storage reliability.
安全性:本文所设计的系统通过用户身份认证等措施保证流媒体服务器、Web服务器与HDFS集群的安全;通过Hadoop的安全机制实现了Web服务器、流媒体服务器和HDFS集群的通信与数据传输安全。
Security: The system designed in the paper ensures the security of streaming media server, web server and HDFS cluster through measures such as user identity authentication. It realizes the security of communication and data transmission of web server, streaming media server and HDFS cluster through the security mechanism of Hadoop.
可扩展性:借助Hadoop软件强大的存储可扩展能力,可以根据需要随时扩展HDFS集群的规模,增加系统的视频文件存储能力。
Expandability: with the help of Hadoop’s strong expandability of storage, the system may expand the scale of HDFS cluster as needed to increase the storage ability of video file.
经济性:HDFS是专门针对廉价计算机设计的分布式文件系统,具有良好的兼容性,任何型号、性能的计算机都可以通过安装该软件加入视频文件存储集群。因此可以充分利用用户现有的廉价硬件资源,节省系统建设投资。
Economical Efficiency: The HDFS is a distributed file system designed specially for cheap computers. It has a good compatibility, and computers of any type or property can join the video file storage cluster by installing the software. In this way, the system can fully use current cheap hardware resource of user and save system cost.