Research On Micro Service Data Platform And Resource Scheduling For Internet Of Things

With the development of information technology,Io T is widely used in various fields.The access of a large number of sensing devices makes the Io T system need to acquire and process large-scale data,and the task of big data computing puts great pressure on the resource-limited Io T system.The traditional Io T data platform is a monolithic architecture,which is easy to develop but cannot provide good computational performance under high concurrency.Besides the monolithic architecture platform has low scalability and is difficult to maintain.It is easy to bring the whole platform down if a module has problems.To solve the above problems,this thesis analyzes the characteristics and requirements of current Io T systems in combination with actual Io T application projects,and proposes an Io T system framework based on edge computing technology,adding edge gateways and edge servers to the system,which can effectively reduce the computational pressure on the cloud and improve the data transmission rate.Under this framework,this thesis researches and develops a reusable Io T data platform based on microservice architecture according to the characteristics of complex data types and large computation quantity of Io T system,and establishes an Io T computation resource scheduling model with computation resource utilization and quality of service Qos as the optimization objectives,and proposes an Io T computation resource scheduling algorithm based on genetic algorithm.The main contents of this paper are as follows.(1)This thesis designs and develops the overall architecture of the Io T-based intelligent pipeline monitoring system,and do buried experimental testing.Combined with the actual application scenario to analyze the current Io T layered architecture and the data platform under the monolithic architecture,the edge computing based Io T architecture is proposed in order to solve the problem that the distance between the measurement nodes and the cloud is far and the big data computing brings huge computing pressure to the system.(2)Under the framework of Io T based on edge computing,the thesis designs and developed.Each microservice of the platform implements only one function,which is independent of each other and has its own storage space.The platform has common functions required by Io T system,which can be used by different application objects,and personalized function modules can be added on the basis of the original platform when necessary.In the Io T application project "intelligent pipeline monitoring system" and "intelligent IOT negative pressure isolation transfer bin" project,the microservices-based Io T data platform is connected for experiments to test the functional integrity and feasibility of the platform,and to compare with the monolithic architecture.The performance of the microservices-based Io T data platform is compared with that of the monolithic Io T data platform for multiple concurrency.The experimental results show that the platform is functionally complete,scalable and reusable,and provides good performance in the case of multi-user concurrency,with a maximum response time difference of 4.5 times from the monolithic architecture.(3)The characteristics of the Io T computing resource scheduling problem are analyzed on the basis of the architecture of this paper,and a mathematical model of Io T computing resource scheduling based on edge computing is established.The model sets up multiple constraints according to the computing resources and bandwidth of the edge server.And a genetic algorithm-based Io T computing resource scheduling algorithm is proposed,which encodes in the way of task-edge server mapping,introduces the idea of self-adaptation to the genetic operator so that it changes according to the chromosome itself under different iterations,and experiments join the chromosome recovery operation.The simulation experiments compare the optimization effect of the scheduling algorithm proposed in this paper and the polling algorithm on the computing resource scheduling model proposed in this paper under different number of tasks.The experimental results verify the effectiveness of the Io T computational resource scheduling algorithm in this paper for the computational resource scheduling problem in Io T systems.