Research On Multi-phase Service Composition And Optimal Scheduling Based On Microservices

In recent years,more and more software exists on the Internet in the form of microservices.The types of services are increasingly diversified and the number of services is growing rapidly.It brings great convenience for developers to shorten the research and development cycle and improve product quality,but at the same time it causes the problem of low efficiency in relying on users to manually find the required services.In order to improve the user’s search efficiency for the required services,improve user satisfaction,and effectively use the service provider’s resources,this paper first defines the construction of the service component library.Based on the service component library,the three-stage research of service demand discovery method,service recommendation scheduling method and service composition scheduling method is carried out,and finally a service that meets user needs and has excellent performance is found.The details are as follows:(1)In order to achieve effective organization of many service resources,this paper defines the construction of a service component library,that is,on the basis of service resources,Service-Cluster and Service-Templet are established based on the characteristics of service similarity,a priori and correlation.Based on word similarity calculation and PI calculation verification,a Service-Cluster or Service-Templet that meets the user’s functional requirements is obtained,which lays the foundation for subsequent research.(2)After the service demand is discovered,the service component library returns to a Service-Cluster.In order to schedule a specific service instance within the candidate service set mapped by the Service-Cluster,this paper will use the factorization machine model to mine the secondary characteristic relationship among the service quality attributes to improve the service Recommendation effect,and construct a hash table of service matching to improve the efficiency of service recommendation,so as to achieve accurate matching based on the user’s service quality requirements and the service quality of the service.(3)After the service demand is discovered,the Service-Templet is obtained.In order to schedule a service instance within each Service-Cluster included in the Service-Templet to combine several service instances to meet user needs,this paper will define the calculation of the combined service quality of service completely,which is used as a standard to measure the quality of combined solutions.Based on swarm intelligence algorithms and machine learning,the combined solutions are solved,and a solution with the best global combined service quality has been obtained.(4)Based on the theoretical research of the above multi-stage service composition and optimized scheduling method,the prototype system of the multi-stage service composition and optimized scheduling is designed and implemented.At the same time,the feasibility of the above theory and method is verified,and the research results are visualized shown.