Traffic Classification Using CNN-LSTM Based on Multimedia Communication System

Network traffic classification is the process of determining and categorizing network data by identifying various kinds of services, applications, and threats. However, the challenge lies in recognizing and classifying the various class of network traffic among high volume of network traffic data. The classification with high network traffic data struggles from imbalanced data and noise. This research proposes an effective Convolution Neural Network - Long Short-Term Memory (CNN-LSTM) model to classify the network traffic based on multimedia communication system. Initially, the proposed CNN-LSTM is evaluated using Information Security centre for Excellence (ISCX) Virtual Private Network (VPN) and non-VPN traffic dataset to determine the model performance. After obtaining data, the min-max normalization is applied to transform numerical values in a standardized range followed by feature extraction using CNN to capture spatial relationships in network patterns. Finally, the proposed CNN-LSTM is performed to classify the network traffic effectively by utilizing CNN for spatial feature extraction and LSTM to learn sequential dependencies of data effectively in traffic flow. While compared to existing methods like Advanced Density-Based Spatial Clustering of Applications with Noise (A-DBSCAN), the proposed method obtains a high accuracy of 0.9876 using ISCX VPN-non-VPN traffic dataset.