A Research On Software-Defined Industrial Internet Of Things Computing Offloading Technology

In recent years,with the rapid development of embedded and mobile devices,the scenes of different devices are becoming more and more abundant,and users are increasingly demanding in different scenarios,and the rise of the Internet of Things is closely related to this.The Industrial Internet of Things(IIoT)is used in the industrial and commercial fields of the Internet of Things.Compared with the general Internet of Things,due to the particularity of the application scenario,the risk of deployment is greater,system failures and downtimes,and applications cannot be real-time.The response can cause unpredictable losses and dangers.At the same time,there are a large number of computational offloading requirements in IIoT.These computational offloading tasks require real-time guarantees,so a reasonable computational offloading architecture and scheduling scheme for IIoT is needed.Based on these requirements,firstly,this thesis summarizes the current status of the integration of IIoT technology and IIoT with Software Defined Network(SDN),explains the problems,and proposes a software-defined industrial Internet of Things based on fog calculation for these problems(Fog-Based Software Defined Industrial Internet of Things,FB-SDIIoT)architecture provides centralized control for computing offloading.At the same time,it analyzes in detail the advantages and technical difficulties brought by the combination of architecture to industrial IoT computing and offloading.Subsequently,the detailed overall architecture design and interface protocol design of FB-SDIIoT was carried out.Starting from the original SDN architecture,the controller agent and the fog node agent software are used to expand the functions to form a new three-layer architecture,and the functions and module design of each layer are elaborated in detail.At the same time,the Fog Node Control and Manage Protocol(FNCAMP)is added on the basis of the OpenFlow standard protocol as the bearer to implement the management and control of the fog node,thereby completing the deployment of the service and offloading for the calculation.The task scheduling in the middle provides the necessary data.Next,the core scheduling module responsible for computing offloading under the FB-SDIIoT architecture is elaborated.Starting from the overall design of the core scheduling module,the module functional requirements and the relationship between the interfaces are explained.The task flow scheduling algorithm(TFSA)is analyzed in the core scheduling module.The task flow scheduling problem is modeled and the corresponding algorithm is proposed for the model.The simulation verifies the algorithm compared with the general algorithm.The advantages of load balancing algorithms.At the same time,the task flow failure scenarios that may occur under the architecture are analyzed.The corresponding Task Flow Adjust Algorithm(TFAA)is proposed for each scenario.The simulation proves that the TFAA algorithm can quickly recover the failed task under light load ensuring the real-time processing of tasks under the architecture.Finally,software implementation and testing of FB-SDIIoT computing offloading system are carried out.The module composition and module relationship of controller side and fog node side agent software are designed and implemented in detail.Combined with FNCAMP protocol,the overall agent software is designed.The state machine that is running.On this basis,the test platform was built by the open source controller system ONOS and the open source switch software OvS.The function of the design and implementation of FB-SDIIoT was tested,and the control and management of the fog node under the platform was verified and the schedule of the uninstall task was calculated.