kQUIC: A Kernel-based QUIC Transport for IoT

Welcome to the kQUIC project! This kernel-based QUIC transport protocol is tailored for IoT environments and was developed in SIOTLAB at Santa Clara University by Puneet Kumar (Ph.D. student) under the guidance of Professor Behnam Dezfouli.

Overview

The kQUIC project enhances the Linux kernel to integrate QUIC as a transport protocol, making it suitable for IoT systems. The implementation provides efficient handling of QUIC connections directly within the kernel, enabling robust and high-performance networking for IoT applications.

---

Features

• QUIC in the Kernel: Implements QUIC as a native transport protocol using IPPROTO_QUIC.
• Seamless Integration: Supports standard socket operations (read, write, msghdr struct association).
• Application Multiplexing: Provides enhanced support for application ID management through ancillary data (struct msghdr *msg and cmag_type fields).

---

Project Structure

main tree:
|
|----include
|       |
|       |-- net
|       |-- uapi
|
|----net
|      |
|      |-- ipv4
|      |-- netfilter

---

Installation

Step 1: Clone the Linux Source

git clone https://github.com/torvalds/linux.git

Step 2: Modify and Compile the Kernel

1. Add the necessary QUIC-related files to the source tree, as per the files/directories mentioned in the source tree.
2. Update the Makefile to include the QUIC implementation.
3. Compile the kernel:

   make -j$(nproc)
   sudo make modules_install
   sudo make install

Step 3: Boot into the New Kernel

Reboot your system and select the newly compiled kernel.

---

Usage

1. Socket Operations:

   • Use IPPROTO_QUIC as the protocol type during socket creation.
   • Socket operations (read and write) function similarly to UDP or TCP sockets.

2. Application Multiplexing:

   • Associate the msghdr struct with the socket APIs to manage application-specific data.

3. Ancillary Data Handling:

   • Use struct msghdr *msg and the cmag_type field to specify the application ID.

---

Contributing

We welcome contributions to enhance and optimize kQUIC! Please submit a pull request or open an issue to discuss your ideas.

---

Acknowledgments

Developed by Puneet Kumar, Santa Clara University, under the mentorship of Professor Behnam Dezfouli at SIOTLAB.

---