Research On Dynamic Multipath Encryption Communication Service For Smart Integration Identifier Networking

The design flaws of traditional network architecture "static and rigid" make it unable to support the high-quality development of Internet economy.In view of the problems existing in the traditional network architecture,the design concept of hierarchical and domain division and intelligent integration is introduced into the Smart Integration Identifier Networking(SINET),which provides a new idea for the innovation of network architecture.With the rapid development of the Internet,the user’s demand for information encryption is increasing.The problem of information encryption under the SINET has gradually become a research hotspot.Therefore,this paper studies the encryption communication service in the context of SINET,aiming to solve the low efficiency of traditional encryption service and improve the reliability of SINET.This paper designs and implements an encryption communication service based on SINET scenario.This service combines the network paths corresponding to core network IPv4,IPv6 and SINET to encrypt and forward user data streams.On this basis,a multi-path data stream scaling algorithm based on deep reinforcement learning is proposed.The main work of this paper is as follows:First,this paper designs an encryption communication service architecture based on SINET,and implements information collection module,information storage module,encryption and transmission decision module.By analyzing the reasons for the low encryption efficiency of existing encryption algorithms,this paper puts forward an algorithm based on matrix to encrypt ack and seq fields in data stream TCP,as well as a transmission strategy that combines core network IPv4,IPv6,SINET multiple network paths.Secondly,this paper discusses the resource consumption of existing secure communication services,designs the level of data flow encryption requirements,and combines the network resource cost to develop a data flow processing scheme based on fuzzy control logic algorithm.Then,this paper analyses the problem of unbalanced proportional distribution of multipath forwarding data streams,completes the problem scenario modeling,defines the state,action space and reward function,designs the internal structure of the neural network,and puts forward a load balancing algorithm for data streams based on deep reinforcement learning.The algorithm collects traffic load and data flow encryption requirements of each path as information basis,and dynamically adjusts the proportion of data flow distribution with the goal of traffic load balancing of each path.In addition,this paper gives the design ideas of load balancing algorithm in control plane and data plane,and provides specific implementation.Finally,an experimental topology is built to test the functionality and performance of encryption communication services.The test results show that the encryption communication service can provide different encryption and forwarding strategies according to the user’s needs,and has lower encryption and decryption time and total transmission time than other encryption algorithms.The load balancing control algorithm has good stability and better traffic load balancing adjustment ability than other control algorithms.