Traceability and Prevention System Construction of Network Attacks Driven by Big Data
Benfa Liu · 2023
The traditional mode of defending against DDoS attacks forms a distributed and strongly coupled system by integrating data processing and control logic into network devices. This system structure can improve the reliability of network operation, but it also has some problems, such as overburdened managers, unable to achieve global unified scheduling and consuming huge resources. In this paper, typical network attacks such as SQL(Structured Query Language) and DDOS (Distributed Denial of Service) are taken as research objects, and the characteristics of these two types of attacks in the context of big data are modeled. The packet is tested by NS2 simulation software, and the results show that BP(Backpropagation) algorithm is used to improve the probabilistic packet marking method, which can effectively improve the convergence speed of the tracking algorithm and improve the tracking accuracy. In view of the large number and high complexity of attack sources in big data environment, it is difficult to solve them by traditional methods based on experience or theory. This paper intends to improve the existing traceability methods by using the adaptability and self-learning ability of BP network. When the number of attack sources is 60, the false alarm rate of the traditional packet marking algorithm is 0.34, and the false alarm rate of the improved probabilistic packet marking algorithm is 0.20.