Research On Real-time And Security Of Smart Factory Network Based On SDN

With the development of a series of information technologies,a new round of industrial revolution has been stimulated,which promotes industrial internet of things and the development of personalized production.Industrial internet of things needs to integrate computing,communication,and control,including environmental awareness,edge computing,high-performance communication,etc.,to provide strong support for personalized production in smart factory.However,the traditional centralized architecture of industrial internet of things is bloated and inefficient,which is gradually unable to match the emerging manufacturing needs.Therefore,the construction of a new type of industrial internet of things for personalized production in smart factory has great importance for promoting new manufacturing services.This paper will try addressing the problems in the traditional industrial internet of things architecture,and put forward the research on the real-time and security of the SDN-based industrial internet of things in the personalized production for smart factory to improve the performance.The main work of this paper is as follows:The paper first explains the current design of the industrial internet of things,and analyzes the related problems of real-time and security.Then,we analyze and deconstruct the original industrial internet of things in smart factory,discuss the requirements and goals of the new industrial internet of things,and propose a Spine-Leaf network topology to replace the Tree network topology.After that,we introduce how the new architecture works in detail.At last,we combine with SDN technology,define the relevant network layer and analyze the overall network layout to perfect the architecture.Secondly,based on the SDN network architecture,the machine learning method is studied,and a route optimization method based on Q-Learning is proposed.We simplify the proposed SDN network architecture and build a mathematical model;then we design the algorithm and define the variables used by the algorithm according to the personalized production settings in smart factory;finally,a related production example is designed to test the effectiveness of multi-dimensional parameter routing optimization method based on Q-Learning.Thirdly,the Blockchain is studied under the SDN network architecture,and we propose a Blockchain method based on the design of Bitcoin.A multi-center partial decentralized system on the industrial edge side is designed to address the security problems caused by a single node.We also make some improvements to enhance data protection.Last but not least,an analysis about the security and privacy of the mentioned method is carried out,and a Blockchain test case based on the proposed architecture is developed to test the effectiveness.Finally,based on the proposed SDN network architecture,we transform an example of a candy automated production platform to test the real-time and security performance of the architecture.The experiments show the rationality and effectiveness of the architecture.