论文部分内容阅读
流媒体数据的分布式处理的研究在流媒体处理领域具有重要的意义,电信网、广播电视网、计算机通信网的不断融合促进了媒体数据的广泛传播,随着数据量的不断增大数据处理的速度变慢,而传统的多处理器并行计算难以满足海量流媒体数据的并发处理需求.使用当前流行的开源分布式处理系统Hadoop对流媒体数据进行分布式计算将大大提高流媒体数据的处理速度,虽然Hadoop常用于大数据企业类业务数据处理,但其对实时数据处理支持较差.因此对Hadoop的调度和负载策略进行研究,通过改进Hadoop的实时调度和负载均衡策略,使其可以满足流媒体处理的实时性、并发性和处理速度的要求.实验结果表明提出的算法和策略使得Hadoop能很好的适应实时流媒体的处理,减少了处理的时间.
The research on the distributed processing of streaming media data is of great significance in the field of streaming media processing. The continuous integration of telecommunication networks, radio and television networks and computer communication networks has promoted the widespread dissemination of media data. With the increasing amount of data, data processing The traditional multiprocessor parallel computing can not meet the demand of massive streaming media data concurrently.Using the popular open source distributed processing system Hadoop to distribute streaming media data will greatly improve the processing speed of streaming media data , Although Hadoop is often used in big data enterprise business data processing, but its support for real-time data processing is poor.Hadoop scheduling and load strategy research, by improving Hadoop’s real-time scheduling and load balancing strategy, it can meet the flow The real-time media processing, concurrency and processing speed requirements.The experimental results show that the proposed algorithm and strategy makes Hadoop well adapt to the real-time streaming media processing, reducing the processing time.