Research On Service Function Chain Orchestration And Optimization Algorithm For Reliability Assurance

Industrial IoT scenarios have a high demand for reliability in the business.To provide flexible provisioning of complex and diverse services,industrial systems are starting to use technologies for network function virtualization and service function chain orchestration.At the same time,virtualization brings additional service reliability issues that affect the business.In this paper,we study the service function chain scheduling and optimization algorithms for service function chain reliability assurance in industrial scenarios.According to the planning and deployment and execution phases of the service,this paper proposes two research points corresponding to the reliability assurance methods in the pre-and postexecution phases of the service,respectively.First,the service function chain orchestration mechanism based on federal reinforcement learning is studied.This method can realize multiple decision makers to collaboratively learn the scheduling experience and obtain the reliability service function chain scheduling results while protecting data privacy.In this paper,we use a federated reinforcement learning model,in which each training participant trains a reinforcement learning model for service reliability improvement in its own environment,and passes the model parameters to the aggregation server after masking them with encryption technology to generate a federated reinforcement learning model,and then sends the federated model back to each privacy domain for model update and parameter correction,so that the decision experience can be shared among different subjects.In this paper,simulation experiments are designed to simulate and analyze the proposed model,and from the results,it can be seen that the proposed model in this paper can achieve service orchestration with reliability guarantee and resource overhead control.Secondly,the reliability assessment and migration strategy of service function chain based on Petri net model is studied.In order to guarantee the reliable execution of the orchestrated services,this paper studies the service function chain reliability optimization method.First,a reliability evaluation model is established and optimization indexes are established.In this paper,we use Petri net model to model the composition and reliability influencing factors of service function chain,simulate the execution process of service,and obtain the reliability evaluation results with time-varying characteristics,which can evaluate the reliability of service more comprehensively.Secondly,the reliability optimization method is studied.The optimization method proposed in this paper is based on the analysis of service reliability,considering the impact of resource sharing on service reliability,designing VNF migration strategy,which can reduce the occupation of extra resources,and also considering the migration cost factor,which can reduce the cost while guaranteeing the reliability.To verify the effect of the algorithm,different scenarios are designed in this paper to evaluate the performance of the optimization algorithm,and the results show that the optimization algorithm proposed in this paper has better performance in improving the reliability of the service function chain and controlling the cost.