Research On Service Dynamic Orchestration Strategy Based On QoS In Internet Of Things

With the deployment and promotion of the application of the IoT,service-oriented architecture has injected new vitality into the development of the IoT.Service computing in the environment of the IoT abstracts and encapsulates the functions of physical devices and provides them in the form of services,thus realizing the sharing and interaction among services of the IoT.However,most of the existing Internet services architecture is extended from Web services architecture,its service orchestration seldom consider factors such as the highly dynamic,massive,heterogeneous,limited resource and energy consumption of IoT services,especially abstract physical device services,the uncertainty of QoS will seriously affect the entire service orchestration,making it difficult to meet the end-to-end QoS constraint requirements of users.Therefore,based on QoS,this thesis focuses on the end-to-end QoS constraints of service orchestration,and studies service orchestration strategies in the IoT environment.Specifically,the thesis mainly completed the following work:1.In view of the dynamic change of QoS of IoT services,the service workflow is divided into three stages: planning,execution and self-adaptation,and a QoS-based service dynamic orchestration strategy is proposed.In the workflow planning stage,the global QoS requirement constraint is decomposed into the local constraint of each workflow node,and the candidate service set of the node is initialized.In the workflow execution stage,the service instances of the candidate service set are evaluated and the optimal service instance is selected for execution.In the self-adaptive stage,the QoS attribute weights of service selection are dynamically adjusted to ensure the end-to-end QoS constraint requirements of the workflow.The thesis strategy and the NSGA-based orchestration strategy are compared and analyzed.The experimental results show that the proposed strategy has better performance in workflow response time,reliability and availability.2.On the basis of 1,the impact of energy consumption factors of IoT services is further analyzed.Use the mapping relationship between services and IoT devices,consider the execution energy consumption of the service and the remaining energy of the physical devices,a dynamic service orchestration strategy based on energy consumption perception is proposed.This strategy builds a QoS-based workflow service energy consumption model,it evaluates the energy consumption of the service instance and associated physical devices when the workflow node selects service instances to balance the service QoS and energy consumption.By comparing with the strategy in 1,the results show that the energy-aware service dynamic orchestration strategy can effectively reduce the energy consumption of workflow execution while ensuring the quality of workflow service,so as to prolong the service life cycle.The research work of this thesis shows that the dynamic service orchestration strategy can effectively guarantee the end-to-end QoS constraints of the service workflow in the IoT environment.On this basis,the improved energy-aware service orchestration strategy can effectively reduce the energy consumption of the IoT application while ensuring the end-to-end QoS constraints.