Research And Application On Service Partition And Selection Strategy In Microservice Platform

With the expansion of the cloud platform business scale,the complexity of the traditional monolithic framework is getting higher and higher,and the maintainability is getting worse and worse.Therefore,the microservice architecture has become a hot research topic.In the current cloud platform,the low service reuse rate and the high code redundancy rate have become a main problem.On the other hand,the current service selection strategies do not comprehensively consider the platform performance characteristics and task characteristics,so these strategies lead to the reduced service execution efficiency.In order to improve the reuse rate of services in the platform and the efficiency of platform execution,firstly,this paper proposes a semantic coupling service partitioning strategy based on domain-driven design ideas.The strategy considers the characteristics of application functions in the micro-service platform and the principle of partitioning of microservices.Secondly,this paper proposes a fine-grained performance prediction model,which can accurately predict the execution time of each microservice.Then,based on the performance prediction model,this paper proposes a performance-aware service path selection strategy,which obtains the optimal service path by initializing the service selection path and dynamic adaptive update,this strategy improves the execution efficiency of the application.Finally,we conduct extensive performance experiment for verifying the service partitioning and selection strategy.The experimental results show that the performance prediction model designed in this paper can accurately predict the service execution time.The proposed method can effectively improve the reuse rate of platform services,reduce the execution time of services,and improve the execution efficiency of applications.