SDNCampus Dataset
Description
The SDNCampus dataset provides a collection of flow statistics data across thirty applications available in a campus and other enterprise networks. It serves as a valuable resource for developing and evaluating deep learning models intended for precisely classifying applications generating traffic in a network. Key Features: The dataset precisely documents and labels thirty applications organized in single file for each application. To ensure inclusivity, a range of applications which are found on conventional computers, smart phones and Internet of Things (IoT) gadgets is represented. Efforts have been made to maintain a balanced distribution of samples across applications classes to enhance model training efficacy and generalization.
Files
Steps to reproduce
The dataset was prepared by setting a lab environment which was connected to a campus network to realized realistic traffic patterns such as delay, jitter and congestion for various hours of a day. Each application was run individually to ensure correct labelling of collected Packet CAPture (PCAP) files. The PCAP files were then process by CICFlowmeter, a robust network traffic flow generator designed to process raw PCAP files or live network traffic into detailed flow records. CICFlowmeter extracts a comprehensive set of features from each traffic flow—such as packet counts, byte counts, flow durations, and timing statistics which are essential for network traffic classification tasks