【摘要】 为了通过内容分配达到负载均衡的目的,对流媒体用户的行为特征进行了分析,并根据用户在点播中对不同片源和片源内部各区段的行为规律,提出访问强度的概念,用来表示与分片相关I/O操作的处理频度,基于此定义给出了一种基于内容分片的负载均衡算法。该算法根据访问强度将节目切片,再将其分配给后台各异质节点。通过仿真实验比较算法在不同访问强度定义下的系统响应时间,验证了算法的有效性,达到了负载均衡的效果,系统性能明显提升。更多还原

【Abstract】 The behavior characteristics of video-on-demand users is used to balance the system content load distribution. The segment access intensity, which indicates the I/O manipulation frequency of the segment, is defined in terms of the inter-movie and intra-movie skewness. A content-skewness-based load balance algorithm was developed to distribute I/O jobs among the heterogeneous nodes. The algorithm uses parallel streaming servers to divide movies into small segments and to then store them across multiple server nodes according to their capabilities. Simulation test results verify the algorithm’s effectiveness with the different definitions of segment access intensity in terms of the mean response time.更多还原