摘要： 为实现网格中视频资源服务的动态负载均衡，对一种动态负载加权均衡算法进行改进。利用监测与发现系统收集每台视频服务器的CPU利用率等主要负载参数，运用上述参数加权得到综合负载，对相邻时刻的负载做平滑处理以避免调度抖动。通过比较平滑后的动态负载值与服务器综合负载阈值进行动态调度，改变相应节点的负载，避免视频服务器间的负载失衡。实验结果表明，该算法能有效降低系统平均服务延迟时间并提高吞吐量，从而提升视频资源网格服务的整体性能。

关键词: 视频资源网格, 监测与发现系统, 负载均衡, 加权负载

Abstract: A method of weighted and smoothed dynamic load balancing algorithm is designed to solve the problem of balancing the load of the video resources. Monitoring and Discovery System(MDS) is used to collect every video server’s parameters such as CPU utilization and these parameters are weighted to get the integration load, adjacent time loads are smoothed to avoid unstable dispatch. Compared with the threshold value of server load, dispatch is executed dynamically by changing the load of corresponding nodes. Experimental results show the algorithm can avoid the imbalance of load in service, reduce the average delayed time in service effectively and improve the throughput.

Key words: video resource grid, Monitoring and Discovery System(MDS), load balancing, weighted load

中图分类号: 

• TP311.52