Research On Content Placement And Cache-aware Scheduling For Streaming Media In Cooperative Cloud-Edge Environment

With the explosive growth of mobile terminal devices and wearable devices,more and more services are provisioned by the cloud,and video has become the world's major mobile data traffic.A large amount of data may cause a large backhaul link burden and a long delay.The traditional centralized cloud processing network architecture has been unable to meet the needs of users.The industry and academia have proposed a network architecture for storing content and implementing application functions at the edge of the network,by proactively storing streaming media content at the edge of the network to achieve low latency acquisition of content.However,due to the limited storage and computing resources of the edge network,how to effectively utilize the limited storage space of the edge network to minimize the overall user response latency has become an urgent problem to be solved.At the same time,the execution time of content placement job in cloud data center can be reduced by making full use of the cache space,which also plays an important role in improving the overall performance of the system.Therefore,it is of great theoretical and practical significance to study the content placement and cache-aware scheduling methods for streaming media under cooperative edge-cloud environment.In view of the above problems,this thesis studies the streaming media cache in cooperative edge-cloud environment:(1)In order to reduce the cost of content placement and the delay of users to obtain placement content in cooperative edge-cloud environment,this thesis designs a content placement based on the latency-cost tradeoff(LCT)algorithm in cooperative edgecloud environment.The algorithm first analyzes the content popularity,the storage and computing power of the edge nodes,and establish a content placement model with the goal of minimizing the latency of content delivery and the cost of data required to place content.In this thesis,the Lagrange multiplier method is used to decompose the problem to be solved into two sub-problems: bandwidth allocation and content placement.Then the content placement problem is transformed into a submodular function optimization problem,and the content placement threshold of the tradeoff between delay and cost is calculated by analyzing the marginal gain of the submodular function.Finally,the proactive content placement algorithm is executed within the threshold range,and the edge node implements the reactive cache replacement algorithm according to the marginal gain decision to achieve the optimal tradeoff between latency and cost.(2)In order to improve the cloud data center processing capability to the streaming media under cooperative edge-cloud environment,this thesis designs a cache-aware scheduling algorithm based on neighborhood search(CANSS)in the cloud data center.The algorithm first divides the jobs into three types,and assigns the resources of the nodes according to the historical information of the nodes executing the jobs.Then,according to the ability of the nodes to perform different types of jobs,the nodes with similar capabilities are clustered based on the neighborhood search.Finally,the job scheduling algorithm is executed based on cache awareness.If the node does not contain the data required for the job,the job scheduling is performed in the result of the node clustering,thereby achieving the goal of shortening the execution time of all the jobs and improving the processing capability of the cloud data center.(3)In this thesis,the proposed algorithm is experimentally verified.In the experiment of content placement algorithm based on the tradeoff between latency and cost in the cooperative edge-cloud environment,cloud-edge cooperative(CEC)cache scheme is compared with edge cache(EC)scheme and cloud-edge non-cooperative(CENC)cache scheme.the average latency and cost of CEC scheme is better than EC and CENC.At the same time,the proposed LCT algorithm in this thesis is compared with HMAX and CMIN algorithm.The average latency and cache hit rate of LCT algorithm are better than CMIN algorithm,and its cost is better than HMAX algorithm.In the experiment of job scheduling algorithm based on cache awareness in cloud data center,CANSS algorithm is compared with CATS algorithm and MDS algorithm.The experimental results show that CANSS algorithm has better job execution time than CATS algorithm and MDS algorithm,and better memory utilization than MDS algorithm.