Research And Implementation Of IoT Microservices Platform Supporting Multiple Protocols

As a key component of IoT industry structure,the IoT platform must not only meet the access requirements of massive heterogeneous devices,but also build efficient and stable interactive channels between devices,users and the cloud.Therefore,the IoT platform needs to meet the expansibility,so that new devices with heterogeneous protocol are able to access the IoT platform.And the IoT platform needs to implement the availability,in order to provide good service quality for the interaction of massive devices and users.Build and deploying the IoT platform with Microservices Architecture and Docker container technology is a feasible solution for maintaining expansibility and availability of the platform.Based on the EMCloud IoT platform of the laboratory research,this paper studies the problems existing in maintaining expansibility and availability of the platform.The work and innovations of this paper are as follows(1)This paper analyzes the requirements of expansibility and availability of the EMCloud IoT platform,proposes the specific problems in the process of construction,deployment and operation of the IoT platform,studies the technology related to the problem,and proposes the design ideas of this paper(2)The IoT Microservices platform supporting multiple protocols is built and deployed in this work.The IoT platform based on Monolithic architecture is divided into a group of microservices,and the appropriate service communication mechanism is designed?so that the platform can meet the expansibility.Aiming at the problem that different services should deploy different number of instances in the IoT platform,a microservice redundancy model based on hierarchical fuzzy comprehensive evaluation method is proposed to evaluate the redundancy of microservices,and microservices will be deployed according to their redundancy.In this way,a microservices cluster will be built to ensure the availability of the platform.(3)A load balancing scheduling strategy based on session identification is proposed in this work.The microservice cluster for the IoT platform needs to meet the requirements of balance,locality and session persistence while dispatching load.This paper proposes the Maglev-Round load balancing scheduling algorithm and implements the session stickiness based on the device session identifier when the network address of device changes.(4)A container dynamic scheduling strategy based on service load and performance prediction is proposed in this work.Aiming at the problem that the number of container instances cannot meet the service requirements because of the change of service load,this paper combines the service load prediction model and the service performance prediction model to predict the service performance status.The container scheduling decision is made according to the predicted value to achieve horizontal scaling of the service.Finally,this paper uses the initial deployment strategy of microservices to deploy the IoT platform,and tests the basic functions of the platform,multi-protocol support features,load balancing scheduling strategy and container dynamic scheduling strategy in order to verify the expansibility and usability of the platform.