Use of the neural network device in network security systems

Автор: Abdurakhmanov R.P.

Журнал: Теория и практика современной науки @modern-j

Рубрика: Основной раздел

Статья в выпуске: 6 (24), 2017 года.

Бесплатный доступ

In this paper, we consider the possibility of using the apparatus of neural networks for the tasks of classifying multimedia traffic in telecommunication networks. For the implementation of a class classifier, a multilayer perceptron was selected. When examining modern attacks at the transport and / or network level of the network, the following characteristics were selected: protocol type, fragmentation characteristics, TTL, ToS, number of sent / received bytes, whether the broadcast, IP options, correctness of the CRC, code and type of ICMP- Messages, TCP sessions-duration, number of flags in the session, whether the window size was changing and others.

Еще

Классификация ip-трафика, classification of ip traffic, p2p-трафик, p2p traffic, evaluation criteria, intrusion detection system, one-class classifier, multilayer perceptron

Короткий адрес: https://sciup.org/140271854

IDR: 140271854

Статья научная