Deep Packet Inspection with cuda GPU

Overview

This project implements a Deep Packet Inspection (DPI) application designed to classify network packets based on custom-defined rules. The application leverages CUDA for high-performance packet processing on the GPU.

The main goal of this project is to build a high-throughput DPI engine that can detect and classify network packets according to specific rule-based conditions. The detection phase is optimized using a tree-based structure to store rules, enabling efficient parallel processing of packets.

(back to top)

Getting Started

This is an example of how you may give instructions on setting up your project locally. To get a local copy up and running follow these simple example steps.

Prerequisites

• NVIDIA GPU with CUDA support (e.g., RTX 4090).
• CUDA Toolkit (version compatible with your GPU).
• Libpcap for reading pcap files.
• CMake for project compilation.
• C++17 or later for building the project.
• yaml-cpp for configuration settings

Installation

2. Clone the repository

   git clone https://github.com/smkh1400/DPI-GPU.git
   cd DPI-GPU

3. Built the project using make

   make -B

(back to top)

Usage

1. Configure the settings in config.yml

   isTimerSet: false
   readPacketMode: offline
   chunkCountLimit: 1572864
   chunkTimeLimit: -1
   threadPerBlock: 32

2. Run the application for a specific pcap file using flag -f

   ./main -f <path_to_pcap_file>

   Or for a directory containing pcap files

   ./main -d <path_to_directory_of_pcap_files>

(back to top)

License

Distributed under the MIT License. See LICENSE.txt for more information.

(back to top)

Contact

Seyed Mohammadreza Khosravian: mz2012kh82@gmail.com

MohammadMehdi Sattari: mamadsadtari83@gmail.com

Project Link: https://github.com/smkh1400/DPI-GPU

(back to top)