AXON: A ROS 2 RMW with Shared-Memory/QUIC Transport and QKD/ML-KEM Key Establishment
This paper introduces AXON, a Rust-based ROS 2 middleware implementation that optimizes performance through shared-memory and QUIC transports while enhancing security via two distinct fail-closed TLS 1.3 configurations: one utilizing hybrid post-quantum key exchange and another leveraging pre-shared keys derived from Quantum Key Distribution (QKD).
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the world of robotics, machines do not think in isolation; they think together. A fleet of autonomous vehicles, a factory arm, or a drone swarm relies on a constant, high-speed conversation between different software parts. To make this possible, engineers use a standard language called a middleware, which acts as a universal translator, allowing code written by different teams to pass messages back and forth. For years, the standard for this translation has been built on a system designed for industrial data distribution, treating every connection between computers as if it were the same kind of network problem, whether the computers are sitting next to each other in the same room or miles apart. However, this one-size-fits-all approach is becoming a bottleneck. As robots move into more complex environments, the way they communicate needs to change. They need a system that knows the difference between a message traveling across a single computer chip and one traveling over a wireless link, and they need that system to be secure against threats that do not just exist today, but will exist in the future, including the potential of quantum computers that could break current encryption.
Researchers at the Universidad de León have built a new system called AXON to solve these problems. Instead of forcing all robot communication through a single, rigid pipe, AXON splits the job based on where the data is going. If two parts of a robot system are running on the same computer, AXON uses a method called shared memory, which lets them exchange information directly through the computer's own memory without the slow, wasteful step of sending data through the network stack. It is like two people in the same room whispering directly to each other rather than shouting through a megaphone. When the data needs to travel between different computers, perhaps from a drone to a control station, AXON switches to a modern, high-speed transport protocol called QUIC. This protocol is designed to handle the messy, unpredictable nature of wireless networks better than older methods, ensuring that messages arrive quickly and in order even if the connection is shaky.
The most significant innovation in AXON, however, is how it handles security. The researchers recognized that robot fleets carry sensitive data, such as maps and camera feeds, and that these systems often operate for years. This creates a risk where an attacker could steal encrypted data today and wait for future technology to decrypt it. To prevent this, AXON offers two distinct ways to lock the connection, both of which are designed to be "fail-closed," meaning the system will simply refuse to connect if the security requirements cannot be met, rather than falling back to a weaker, unsafe method. The first option uses a hybrid key exchange that combines a standard mathematical lock with a new, quantum-resistant lock. This ensures that even if a future quantum computer breaks the old lock, the new one will still hold. The second option is even more robust: it uses keys generated by a quantum key distribution system, which relies on the laws of physics rather than math to create unbreakable codes. In this mode, the system imports a secret key provided by a secure hardware device and uses it to lock the connection, with no mathematical negotiation allowed that could be intercepted.
To make this work, the team had to build a new architecture from the ground up. The system is split into two layers: a fast, safe core written in a modern programming language called Rust, and a thin adapter that speaks the standard language of robot software. A small background program, or daemon, runs alongside the robot to discover other machines and manage the map of who is talking to whom. This separation allows the data to flow smoothly while the control system handles the complex task of setting up secure connections. The researchers also had to modify the underlying security software they used, because the standard version did not allow them to use the specific type of quantum-safe keys they needed. They created a specialized version of the security library that allows the system to import a pre-shared secret key directly, bypassing the usual handshake process that could be vulnerable to future attacks.
The results show a system that is tailored to how robots actually operate. By using shared memory for local tasks, the system avoids unnecessary delays. By using QUIC for remote tasks, it handles network errors gracefully. And by enforcing strict security rules, it ensures that the connection is either perfectly secure or does not exist at all. The researchers tested the system in simulations and confirmed that it can handle the flow of data between different parts of a robot system, managing everything from simple status updates to complex sensor data. They also demonstrated that the system can successfully switch between the two security modes, rejecting any attempt to connect with a peer that does not meet the strict security criteria. While the system has not yet been tested against a real, physical quantum key distribution device, and while it is not yet a perfect replacement for every existing robot communication standard, it proves that a more flexible, secure, and efficient approach is possible. The work highlights that as robots become more integrated into our lives, the software that connects them must evolve to be as smart and adaptable as the machines themselves, protecting their conversations not just from today's hackers, but from the threats of tomorrow.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.