The Research And Design Of Energy-saving Strategy In Distributed Streaming Media Monitoring System

In order to reach the great goal of building a safe China,the security video surveillance business has developed rapidly in recent years,the sharpness of video surveillance is getting higher and higher,and the coverage rate of city surveillance is increasing.In order to meet the needs above,the popularity and scale of distributed streaming media monitoring systems are increasing.Although distributed streaming media services to a certain extent solve the performance bottleneck of video surveillance systems,their power consumption is a problem that cannot be ignored.The scale and number of distributed streaming media monitoring systems are constantly increasing,and their growth rate will further increase in the foreseeable future.Then there will be a large amount of energy consumed by countless continuously operating cameras.Transmission and storage will also lead to large energy consumption.Therefore,how to improve the overall energy utilization of the monitoring system without sacrificing video quality is a hot issue that needs urgent solution.The solution of this problem is not only conducive to energy conservation,but also conducive to the further development of distributed systems.In order to solve the problems above,this paper designs a green distributed streaming media architecture that greatly saves the overall energy consumption of the video surveillance system.This paper introduces the green server cluster technology,redesigns the system load distribution strategy combining the load migration strategy,and carries out multi-node collaborative design to improve the overall energy efficiency of the distributed video surveillance system according to the characteristics of the video surveillance system.In addition,this paper also proposes a concept of overlay node group and copy concentration policy to reduce the energy consumption in HDFS,and combines them to propose a green server cluster architecture of streaming media video surveillance system.The green distributed streaming media system designed in this paper contains two aspects: load task distribution and migration and data storage concentration.By using task access algorithm and load migration algorithm,the number of operating nodes in media server cluster is reduced,and some low load nodes in HDFS are suspended through concentrating the HDFS system replica.This archetecture greatly reduces system energy consumption and improves energy efficiency.The main work and achievements of this paper are as follows:1.For the energy consumption of the green streaming media monitoring system,the four states of the node(full load state,working state,standby state,suspend state)are proposed for power management.2.Aiming at the improvement of the energy efficiency of the streaming media monitoring system,this paper proposes a load concentration strategy based on dynamic feedback so that the media distribution server can concentrate the media distribution tasks as much as possible and migrate the tasks of the low load nodes to orther nodes to raise the sleep probability of the low load nodes.3.In order to improve the energy efficiency of HDFS in streaming media monitoring systems,a group of overlay nodes and a centralized replication algorithm are proposed to enable clusters to suspend some data nodes under low load conditions without affecting performance.4.The overall structure and specific modules of the media server cluster of the green streaming media monitoring system were designed,and its main functional modules were implemented.