Research On Scheduling Methods Of Microservices In The Internet Of Things Environment

In the field of cloud computing,service scheduling is an important branch.If the service performance is affected,problems such as service delay,service unresponsiveness,and service feedback error messages may occur,which is why the scheduling problem has become one of the key research problems in cloud computing.In the Io T area,infrastructure environment supporting services has become more complex and factors affecting scheduling have increased,for instance,data distribution is more decentralized and dependencies between services are more complex,which have put forward higher requirements for scheduling.Microservices are a new type of application software architecture.Since each microservice can have its own storage and processing resources,it has advantages such as flexibility and scalability,and has been commonly used in various fields of applications.Therefore,in this paper,the scheduling method of microservices is studied in the context of Io T environment.Firstly,a multi-tier monitoring architecture for microservices in the Io T environment is designed by analyzing the monitoring requirements of microservices in the Io T environment.Second,the scheduling model in the Io T environment is discussed,including the base environment,scheduling constraints,and scheduling goals,and a service scheduling method based on Deep Deterministic Policy Gradient is proposed.Finally,a visual monitoring system is implemented based on a microservice-oriented multi-tier monitoring architecture,and the effectiveness of the Deep Deterministic Policy Gradient-based service scheduling method is also verified in this paper.The main research work of this paper is as follows:1.To solve the demand for comprehensive monitoring of microservice operation status raised by scheduling in the Io T environment,a multi-level monitoring architecture for microservices is designed in this paper.The architecture combines Kubernetes and Istio technologies to realize microservice multi-tier monitoring at the node level,as well as at the container and service levels,from both resource and service perspectives.In particular,the architecture of microservice-oriented multi-level monitoring consists of a service cluster,which is the target of monitoring,and a monitoring server,collecting various types of monitoring data and providing an interface to access monitoring data from outside,and finally,based on this architecture,a visual monitoring server is implemented,which can monitor the operation status of nodes,containers and services and obtain monitoring metrics that can be used to support the execution of scheduling methods.2.In order to meet the changes of factors affecting the quality of microservices in the Io T environment,a scheduling model in the Io T environment is analyzed and designed in this paper,and a service scheduling method based on a deep deterministic policy gradient is proposed.In particular,the service scheduling model designed in this paper takes service dependency,data dependency and user dependency as scheduling constraints,and takes resource balance and service delay as scheduling objectives from the service provider and user perspectives.The service scheduling method proposed in this paper combines deep learning and reinforcement learning to achieve continuous optimization of the scheduling scheme.3.In order to verify the feasibility of the visual monitoring system and the effectiveness of the service scheduling method based on Deep Deterministic Policy Gradient,the surveillance data obtained from the visual monitoring system is processed and used as the experimental data set in this paper,and the proposed method is compared with the scheduling method based on Particle Swarm Optimization(PSO)and the scheduling method based on Genetic Algorithm(GA).The experimental results show that compared with the Particle Swarm Optimization-based scheduling method and the Genetic Algorithm-based scheduling method,the Deep Deterministic Policy Gradient-based scheduling method is able to generate scheduling schemes that better meet the scheduling objectives under the constraints of the Io T environment.The Deep Deterministic Policy Gradient-based scheduling method can generate a scheduling scheme that better meets the scheduling objectives under the constraints of Io T environment.