BGP Anomali Tespitinde Hibrit Model Yaklaşimi

Abstract

Border Gateway Protocol (BGP) is important for the quality of the connection between autonomous systems and the domains it is connected to. With attacks made at this level, any anomaly in the network will cause connection failures at the border gateways. In this study, a classification model is proposed by using machine learning and deep learning algorithms for the detection of BGP anomalies. The proposed model is developed based on decision trees and random forest and multilayer perceptron algorithms. Indirect BGP anomalies and connection failure anomalies in the model were evaluated with accuracy and F1-score. In the tests performed on the Slammer dataset, it was seen that the best result was obtained with 99,47 accuracy, and 98,85 F1-Score value in the model studied with the Hybrit Model. © 2022 IEEE.