Design And Implementation Of Edge Oriented Microservice Path Selection And Resource Scheduling System

With the development of the Internet of things,higher requirements are put forward for the real-time,security and reliability of business.Therefore,complex Internet of things services are decomposed into loosely coupled fine-grained microservices through the microservice architecture to achieve more efficient resource utilization and flexible deployment.Relying on lightweight container technology,edge computing deploys microservice container instances near the edge of users to reduce business communication delay and improve security.The concurrent request of the business will select the container instance for each microservice to form the service execution path,in order to complete within the agreed delay constraint.However,due to the distributed characteristics of edge nodes and the complex workflow architecture of microservice applications,how to select container instances for dependent microservices in concurrent requests to reduce delay and network resource consumption has become a difficult problem.In addition,edge nodes are usually resource constrained,and microservices will compete for edge resources.It is necessary to consider how to schedule resource to expansion container instances,in order to reduce the waiting time of key microservices.Therefore,the research on edge oriented microservice path selection and resource scheduling is of great significance to improve the delay and completion quality of Internet of things services.Most of the existing edge service selection strategies only consider the requests with linear simple structure and ignore the interdependence of microservices in the form of complex workflow.The selection of container instances with a long network distance from each other will lead to a large communication delay and network resource consumption.In addition,concurrent user requests will compete for service instances,Unreasonable service selection will produce unnecessary waiting delay.Existing schemes are difficult to meet this demand.Aiming at how to select the service path of concurrent requests in the edge environment to ensure the service delay and network resource consumption,this paper proposes an edge oriented microservice path selection mechanism.Firstly,a three-tier architecture of microservice selection based on container is proposed.Secondly,considering the dependence between microservices,the request delay model and network resource consumption model are established,and transformed this problem into a multi-objective optimization problem of minimum average delay and network resource consumption under multiple constraints.Then an improved ant colony algorithm based on priority is proposed,which uses the priority mechanism to optimize the task scheduling order,and uses the convergence of ant colony algorithm to obtain the global optimal result.The simulation results show that the microservice path selection mechanism proposed in this paper can limit the delay of concurrent requests and reduce the consumption of network resources.In order to verify the actual effect of micro service path selection and scheduling mechanism,this paper implements an edge oriented micro service path selection and resource scheduling system,designs and implements modules such as resource monitoring,application management,container scheduling and microservice path selection.speak specifically,creates multiple virtual machines to simulates the edge computing environment and establish network topology,and uses Redis and MySQL to store system data.The management of container life cycle is realized by using the container cluster management tool Kubernetes,and the service path selection is realized based on the edge oriented microservice path selection mechanism.Finally,each module is tested from the function aspect,and the actual operation effect of each functionis shown.The test results are consistent with the expected results.