Research On 5G Network Slicing Security Of Distribution Internet Of Things

With the development of communication technology,5G technology has been widely adopted in various industries,including the distribution Internet of Things.In the distribution Internet of Things,different types of applications have different network requirements.In order to meet various differentiated service requirements in the distribution Internet of Things,the network slicing technology in 5G has been applied.However,in the distribution Internet of Things,network slicing creates security risks during its allocation and deployment process.It’s important to study the security of network slices in the distribution network of things because they can seriously affect the security and stability of power systems.This dissertation focuses on the following research areas:Firstly,this dissertation proposes a scheme for constructing fine-grained network slices based on specific application types in the distribution Internet of Things.The essence of this scheme is to divide the network traffic of the same application type onto the same chip.Compared to the traditional method of allocating network slices based on application scenarios,this scheme improves the flexibility and reliability of the distribution network.Secondly,this dissertation proposes a dual traffic classification model based on CNN and self attention mechanism to securely allocate fine grained network slices.The model consists of two parts:a DDoS traffic classification model and a normal traffic multi classification model.Network traffic firstly passes through a DDoS traffic classification model to isolate malicious traffic;then,the normal traffic multi-classification model classifies the normal traffic according to different application types.Experimental results show that this model has a higher classification accuracy compared to other single models,and can be effectively applied to the security allocation task of fine-grained network slices.Finally,this dissertation proposes a new security deployment method for power distribution IoT slices.This method quantitatively analyzes the security of nodes involved in various services in the power grid,proposes a TOPSIS analysis method based on entropy weight method to rank the security values of nodes,and proposes a discrete particle swarm optimization algorithm based on genetic algorithm to deploy virtual nodes and virtual links.During the deployment process,the method in this article also prioritizes the deployment and protection of nodes with high security values.Through a large number of comparative simulation experiments with other deployment methods,it is proved that this method can reduce deployment costs,improve security benefits,and deployment profitability.