基于SNMP和神经网络的DDoS攻击检测

DDoS（Distributed Denial of Service）已经严重威胁计算机网络安全。对DDoS攻击检测的关键是找到能反映攻击流和正常流区别的特征，设计简单高效的算法，实时检测。通过对攻击特点的分析，总结出15个基于SNMP（Simple Network Management Protocol）的检测特征。利用BP神经网络高效的计算性能，设计了基于SNMP和神经网络的DDoS攻击检测模型，提高了检测实时性和准确性。实验表明：该检测模型对多种DDoS攻击都具有很好的检测效果。

DDoS Attack Detection Based on SNMP and Neural Network

Distributed Denial of service （DDoS） attack is a major threat to the security of computer network. It is a challenge to detecting DDoS attack in real-time. The key is to find the difference between normal stream and attack stream and identify it with simple algorithm. By analyzing characteristics of DDoS attack, 15 detecting signatures based on SNMP are summarized. A detecting model, which takes advantage of BP neural network calculation performance in intrusion detection, is proposed in this paper. And computer simulations show that this model could effectively detect many kinds of DDoS attacks.