Research On Network Packet Detection Method Based On Elliptic Curve And Machine Learning

With the introduction of computer technology,network attacks become more and more frequent.Network message is the format of network data transmission.Network message without authentication and detection may contain network attack.Elliptic curve digital signature algorithm(ECDSA)can detect whether the data sender is credible at the physical layer.This paper analyzes the ECDSA algorithm,and designs the hardware implementation based on FPGA according to the analysis of the algorithm.In this paper,the dot multiplication module is designed with the structure of dual-mode multiplier to improve the signature speed.Although the signature algorithm can ensure that the sender is difficult to impersonate illegally,but the threat from the communication channel is still very much,some illegal molecules may attack into the computer through the network to tamper with the message,transmission and other disruptive behavior,sensitive to personal information and industrial control network,the trading system and other havoc.Intrusion detection system is a means to resist network attack.Traditional intrusion detection systems are useful,but do not have the ability to detect unknown attacks.This article turns to machine learning algorithms to detect intrusions,using known attacks to help detect intrusions with similar properties.The experimental results show that the CART-based bagging method provides better accuracy,lower false positive rate and faster model training speed.