Network Anomaly Detection With Temporal Convolutional Network and U-Net Model

Abstract

Anomaly detection in network traffic is one of the key techniques to ensure security in future networks. Today, the importance of this topic is even higher, since the network traffic is growing and there is a need to have smart algorithms, which can automatically adapt to new network conditions, detect threats and recognize the type of the possible network attack. Nowadays, there are a lot of different approaches, some of them have reached relatively sufficient accuracy. However, the majority of works are being tested on old datasets, which do not reflect current network conditions and it leads to overfitted results. This is caused by high redundancy of the data and because they fail to reflect the performance of the latest methods in the real-world anomaly detection applications. In this work, we applied a couple of new methods based on convolutional neural networks: U-Net based and Temporal convolutional network based for network attack classification. We trained and evaluated methods on the old dataset KDD99 and the modern large-scale one CSE-CIC-IDS2018. According to results, Temporal convolutional network with LSTM has achieved accuracy 92% and 97% on the KDD99 and the CSE-CIC-IDS2018 respectively, the U-Net model has accuracy 93% and 94% on the KDD99 and the CSE-CIC-IDS2018 respectively. Additionally, we utilized the focal loss function in the Temporal convolutional network with Long Short-Term Memory model, which has positive effect on class imbalance in time-series data. We showed, that the Temporal convolutional network in combination with Long Short-Term Memory network and U-Net model can give higher accuracy compared to other network architectures for network traffic classification. In this work we also proved, that methods trained on the old dataset can easily overfit during training and achieve relatively good results on the testing set, but at the same time, these methods are not so successful on more complex and actual data.

Network Anomaly Intrusion Detection Based on Deep Learning Approach

“…Although the accuracy of [ 12 ] is 99.7% better than our paper, the paper does not use all datasets for training and testing, so it cannot be compared with our paper. Compared with [ 5 , 6 , 9 , 10 , 17 ], the proposed model effectively improves the detection performance.…”

“…For example, the authors of [ 26 ] randomly selected 40,000 benign data (the total number of benign traffic data was 13,484,708) and 20,000 attack data to conduct experiments. The authors of [ 6 ] used nine of the ten files for their experiments. In this study, we used all datasets for experimental evaluation.…”

“…In Ref. [ 6 ], Mezina et al used the KDD99 and CSE-CIC-IDS2018 datasets and applied a U-Net and temporal convolutional network (TCN) to identify network attacks. They combined a TCN with LSTM to achieve a higher accuracy; the TCN mainly solved time series data by adding dilated convolution, which was able to increase the scope of convolution.…”

“…Accordingly, research has been conducted on publicly available network intrusion detection datasets; the most commonly researched datasets have been KDD Cup 1999 (KDD99) and NSL-KDD [ 5 ]. The network traffic of these two datasets have been sufficient to detect intrusion-spreading viruses, but today’s attack methods have diversified so these datasets are outdated and unreliable [ 6 ]. The CSE-CIC-IDS2018 dataset is derived from real network traffic data, and it can be applied to actual network detection capabilities through deep learning methods [ 7 , 8 ].…”

A Lightweight Convolutional Neural Network based network intrusion detection and classification method for Social Internet of Things

“…For anomaly detection, He et al suggested combining a Multimodal Deep Auto Encoder (MDAE) and an LSTM [34]. UNSW-NB15, NSL-KDD, and CICIDS2017 datasets are used to test this exclusive approach, and accuracy scores for multiclass classi cation were 98.60%, 86.20%, and 80.20% respectively.Mezina et al employed a U-Net and a temporal CNN trained onCICIDS2018 and KDD99 datasets to identify network threats [35]…”