Research On Data Placement Strategy For Cloud-Side Collaboration Under The 5G Background

With the rapid development of 5G technology,the cloud-edge collaboration model has been proposed by the industry as a new data processing method.It combines the powerful computing and storage capabilities of the cloud with the ultralow-speed edge computing features to complete the edge-side participation.With the development of cloud applications,the cloud supports the collaborative work goal of the edge and the nearest needs,but also has new requirements for the existing data storage mode.How to reduce the user's data access delay and maintain the cloud The balance of data storage with the edge side is a more important issue.Therefore,this paper makes the following researches on the above issues and draws the following conclusions:(1)Aiming at the data storage problem in the cloud data center,with the access delay and the load balance between the internal servers of the cloud data center as the constraints,and the data storage cost as the optimization goal,a data balanced placement strategy is proposed.(2)Combined with the existing research,based on 5G network model,and the research problem is defined.On this basis,a multi-copy data storage model for cloudedge collaboration is proposed.Load balancing between cloud data centers and between edge servers and user access delay are constraints,and a service cost optimization model is proposed,which is abstracted into a multi-objective constraint optimization problem.(3)A multi-swarm collaborative particle swarm optimization algorithm is proposed,and a multi-copy cloud data placement strategy is proposed based on the algorithm,and a program is written in JAVA language for simulation experiments.At the same time,the data placement strategy proposed in this paper is compared with the other two data placement strategies A comparative analysis was carried out on a real dataset.The simulation results show that,on the premise that the access delay of99% of users is less than 150 ms and the delay of some users is less than 30 ms,the average service cost calculated by the data placement strategy proposed in this paper is compared with the data placement based on the TOPR algorithm.The strategy and the data placement strategy based on the random algorithm are reduced by about 24%and 58%,respectively,and under the premise that the access delay of 90% of users is less than 150 ms and the delay of some users is less than 30 ms,the reduction is about28% and 65%.The superiority of the data placement strategy proposed in this paper is verified.