Research And Implementation Of Multi-domain IoT Service Provision Architecture And Combination Method

The Internet of Things is a strategic emerging industry of the country and is widely used in industrial control,smart cities,military defense and other fields.Distributed cross-domain service provisioning is the foundation for agile development and widespread deployment of IoT applications.For the IoT service,it presents the characteristics of diversification,cross-domain,and dynamic reconstruction.This thesis focuses on multi-domain IoT service provision architecture and flexible service combination method,and designs and develops multi-domain micro-objective shared open system to support the development of Internet of Things applications.For the characteristics and needs of the IoT heterogeneous mass service,layered and domain deployment,and application cross-domain collaboration,A multi-domain IoT flat service provisioning architecture based on MSA(Microservice Architecture)is proposed to realize massive dynamic differentiated service state dynamic awareness and service distributed cross-domain management.Service delivery architecture is compatible with cross-domain differentiated IoT devices and diverse information system services;it provides multi-style services at the micro-service level,service group level,and system level,enabling independent deployment of massive services,with loose coupling and high fault tolerance,providing support for cross-domain collaboration.Aiming at the characteristics and requirements of IoT application demand differentiation and service composition diversification,this thesis studies the relationship between multi-domain IoT mass service and proposes a multi-domain massive service relationship map construction and evolution model.A multi-dimensional cross-domain service dynamic discovery mechanism and a demand-driven scenario adaptive service flexible combination method are proposed to meet the differentiated needs of users,shorten the application online cycle,reduce the development and deployment costs,and realize the provision of IoT services on demand.The analysis and verification show that the average recall rate of multi-dimensional dynamic discovery method is increased by 9.59% compared with the method of similarity matching based on service IO;When the service set size is greater than 600,the average time of the demand-driven scenario adaptive service flexible combination method is reduced by 11.5% compared to the automated service combination method that satisfies Top-K.Based on the described techniques and methods,a multi-domain micro-objective sharing open system is designed and implemented.The system realizes real-time sensing of dynamic resources,and the customized service is provided on demand.In the case of distributed cache,the average response time of the service can be up to 46 ms,and in the case of no distributed cache,the processing time is kept from 200 ms to 300 ms,a single node can support up to 500 or so concurrency.The system application case shows such as the campus environment monitoring App and the Fubao smart campus show the availability and effectiveness of the system,and verify the achievability of the architecture and method.