Research On Internet Of Things Platform And Resource Scheduling Based On Hadoop

Iot complex scenarios and multi-source heterogeneous devices access to traditional Internet platform brings a series of information processing and resource scheduling problem,embodied in a large amount of data traffic storage,computing performance is not enough,difficult multi-source heterogeneous data integration,extension reusability is poor,and server resource scheduling algorithm to optimize the effect not beautiful.Therefore,it is urgent to reasonably integrate computing resources of the Internet of Things platform and reasonably schedule specific tasks to improve the performance of the platform.To solve above problems,based on the multi-source heterogeneous data parsing,platform,extensibility and reuse resource scheduling are studied,and set up the high concurrency performance,and have a large number of heterogeneous protocol data processing ability of the Internet of things platform,and designed and developed high reusability,portability and scalability of service module,and puts forward a iot of resource scheduling optimization solution.The research work of this thesis mainly includes three aspects:(1)Aiming at the poor scalability,portability and reusability of the traditional Internet of Things platform,a set of micro-service-oriented Hadoop Internet of Things platform was designed and developed on the basis of the "edge + cloud" Internet of Things architecture.This thesis designed and developed a number of micro-service modules,mainly including user management,device management,device access and multi-source heterogeneous protocol analysis modules,to improve the compatibility of the Internet of Things platform.The "Intelligent pipeline monitoring system" and "Intelligent iot negative pressure isolation transfer warehouse",two successful access platforms of the Internet of Things project,were tested,and the expansibility,portability and reusability of the micro-service module were successfully verified.Compared with traditional iot platforms,the response time is reduced by 42% and the throughput is 1.6 times of the original.(2)Aiming at the problem of Internet of Things platform server dealing with the analysis of multi-source heterogeneous data of Internet of Things,this thesis designs and develops a micro-service module for multi-source heterogeneous protocol.The platform is compatible with multiple Internet of Things communication protocols,and can realize the fusion transformation of HTTP,MQTT,CoAP and custom protocols.The test results prove that the platform has the ability to be compatible with multiple heterogeneous protocol data.(3)Aiming at the scheduling problem between computing resources of the Internet of Things platform and huge tasks,the physical and mathematical model of resource scheduling in the Internet of Things system is established,and the particle swarm optimization algorithm is improved.In this thesis,the iterative updating formula of particle swarm optimization algorithm is changed into an improved algorithm with inertia weight and shrinkage factor to reduce the task completion time.The simulation results show that in the case of 500 tasks,the execution time of the improved algorithm in this thesis is reduced to 75% compared with the original RR scheduling algorithm.Meanwhile,compared with the other three particle swarm optimization algorithms,the improved algorithm has shorter execution time and better convergence effect.The micro-service oriented Hadoop Internet of Things platform developed in this thesis has the characteristics of availability,scalability and concurrency.The platform can deal with the analysis,storage and calculation of a large number of heterogeneous protocol data.On the other hand,the resource scheduling method proposed in this thesis can further optimize the platform execution efficiency.