arXiv:1903.09778 · ACM Special Interest Group on Data Communication
A link layer protocol for quantum networks
Axel Dahlberg, Matthew Skrzypczyk, Tim J. Coopmans, Leon Wubben, Filip Rozpędek, Matteo Pompili, Arian Stolk, Przemysław Pawełczak, Robert Knegjens, Julio de Oliveira Filho, Ronald Hanson, Stephanie Wehner
Quantum communication brings radically new capabilities that are provably impossible to attain in any classical network. Here, we take the first step from a physics experiment to a quantum internet system. We propose a functional allocation of a quantum network stack, and construct the first physical and link layer protocols that turn ad-hoc physics experiments producing heralded entanglement between quantum processors into a well-defined and robust service. This lays the groundwork for designing and implementing scalable control and application protocols in platform-independent software. To design our protocol, we identify use cases, as well as fundamental and technological design considerations of quantum network hardware, illustrated by considering the state-of-the-art quantum processor platform available to us (Nitrogen-Vacancy (NV) centers in diamond). Using a purpose built discrete-event simulator for quantum networks, we examine the robustness and performance of our protocol using extensive simulations on a supercomputing cluster. We perform a full implementation of our protocol in our simulator, where we successfully validate the physical simulation model against data gathered from the NV hardware. We first observe that our protocol is robust even in a regime of exaggerated losses of classical control messages with only little impact on the performance of the system. We proceed to study the performance of our protocols for 169 distinct simulation scenarios, including trade-offs between traditional performance metrics such as throughput, and the quality of entanglement. Finally, we initiate the study of quantum network scheduling strategies to optimize protocol performance for different use cases.
Quantum communication brings radically new capabilities that are provably impossible to attain in any classical network. Here, we take the first step from a physics experiment to a quantum internet system. We propose a functional allocation of a quantum network stack, and construct the first physical and link layer protocols that turn ad-hoc physics experiments producing heralded entanglement between quantum processors into a well-defined and robust service. This lays the groundwork for designing and implementing scalable control and application protocols in platform-independent software. To design our protocol, we identify use cases, as well as fundamental and technological design considerations of quantum network hardware, illustrated by considering the state-of-the-art quantum processor platform available to us (Nitrogen-Vacancy (NV) centers in diamond). Using a purpose built discrete-event simulator for quantum networks, we examine the robustness and performance of our protocol using extensive simulations on a supercomputing cluster. We perform a full implementation of our protocol in our simulator, where we successfully validate the physical simulation model against data gathered from the NV hardware. We first observe that our protocol is robust even in a regime of exaggerated losses of classical control messages with only little impact on the performance of the system. We proceed to study the performance of our protocols for 169 distinct simulation scenarios, including trade-offs between traditional performance metrics such as throughput, and the quality of entanglement. Finally, we initiate the study of quantum network scheduling strategies to optimize protocol performance for different use cases.
Quantum communication brings radically new capabilities that are provably impossible to attain in any classical network. Here, we take the first step from a physics experiment to a fully fledged quantum internet system. We propose a functional allocation of a quantum network stack and construct the first physical and link layer protocols that turn ad-hoc physics experiments producing heralded entanglement between quantum processors into a well-defined and robust service. This lays the groundwork for designing and implementing scalable control and application protocols in platform-independent software. To design our protocol, we identify use cases, as well as fundamental and technological design considerations of quantum network hardware, illustrated by considering the state-of-the-art quantum processor platform available to us (Nitrogen-Vacancy (NV) centers in diamond). sing a purpose built discrete-event simulator for quantum networks, we examine the robustness and performance of our protocol using extensive simulations on a supercomputing cluster. We perform a full implementation of our protocol, where we successfully validate the physical simulation model against data gathered from the NV hardware. We first observe that our protocol is robust even in a regime of exaggerated losses of classical control messages with only little impact on the performance of the system. We proceed to study the performance of our protocols for 169 distinct simulation scenarios, including tradeoffs between traditional performance metrics such as throughput and the quality of entanglement. Finally, we initiate the study of quantum network scheduling strategies to optimize protocol performance for different use cases.
Quantum mechanics shows that if one is able to generate and manipulate entanglement over a distance, one is able to perform certain tasks which are impossible using only classical communication. Classical communication refers to what is used in the Internet of today. A quantum internet would therefore bring new capabilities to our highly connected world. These capabilities both involve (1) the ability to perform tasks with are provably impossible in the current Internet, such as unconditionally secure communication, and (2) the ability to perform certain tasks much more efficient, such as distributed (quantum) computing or extending the baseline of telescopes. To be able to build a quantum internet, two main components are needed: (i) hardware that can store, manipulate and entangle qubits and (ii) a software stack to control the hardware. The core task of both of these is to generate entanglement to be used by applications. In this thesis we focus on the latter, i.e. the development of software and protocols that enable entanglement generation using capable hardware. To enable a certain application, one can certainly, in theory, manually specify each operation the hardware should perform, involving micro-wave pulses, lasers etc. However, in practice this is not feasible, if not to say impossible, due to the complexity of the operations needed, especially in a distributed system such as a quantum network. What is needed is a software stack, which can help with abstracting complexity away in multiple layers. This allows for someone to program a protocol in one layer without knowing all the details of the lower layers. In particular, one can abstract away the hardware details, in order to make higher-layer protocols and applications hardware-agnostic. Therefore, to be able to build a universal, efficient and scalable quantum internet, a software stack is crucial. In chapter 2 we start discussing the networking part of a software stack. Namely, we introduce a network stack for a quantum internet, drawing parallels to the IP/TCP-suite of the classical Internet. We continue with proposing a service and interface of the lowest layer of the network stack: the link layer. The link layer is here responsible for generating entanglement between nodes in a quantum network which are directly connected by a quantum link, i.e. a fiber cable. When developing a protocol or application it is very useful to be able to run it. Both to see if the intended ideas make sense and also to check that the implementation is actually correct. Currently we do not have quantum hardware that exposes a full-fledge API that can be used to execute applications. For this reason, it is very useful to be able to instead simulate the hardware in a way that exposes the same API as the hardware being developed. In chapter 3 we introduce SimulaQron for this exact purpose. Any application of a quantum internet will need entanglement in one way or another. However, entanglement is generally hard to generate and is usually the bottleneck when executing an application. We would therefore like to make use of the generated entanglement in the most optimal way. To be able to do this we need to understand how entanglement can be transformed and distributed in a quantum network. We study the entanglement of a particular class of states called graph states in chapters 4 to 9 and how these states can be transformed in a quantum network.
Abstract Scaling current quantum communication demonstrations to a large-scale quantum network will require not only advancements in quantum hardware capabilities, but also robust control of such devices to bridge the gap in user demand. Moreover, the abstraction of tasks and services offered by the quantum network should enable platform-independent applications to be executed without the knowledge of the underlying physical implementation. Here we experimentally demonstrate, using remote solid-state quantum network nodes, a link layer, and a physical layer protocol for entanglement-based quantum networks. The link layer abstracts the physical-layer entanglement attempts into a robust, platform-independent entanglement delivery service. The system is used to run full state tomography of the delivered entangled states, as well as preparation of a remote qubit state on a server by its client. Our results mark a clear transition from physics experiments to quantum communication systems, which will enable the development and testing of components of future quantum networks.
When built, quantum repeater networks will require classical network protocols to control the quantum operations. However, existing work on repeaters has focused on the quantum operations themselves, with less attention paid to the contents, semantics, ordering and reliability of the classical control messages. In this work we define and describe our implementation of the classical control protocols. The state machines and packet sequences for the three protocol layers are presented, and operation confirmed by running the protocols over simulations of the physical network. We also show that proper management of the resources in a bottleneck link allows the aggregate throughput of two end-to-end flows to substantially exceed that of a single flow. Our layered architectural framework will support independent evolution of the separate protocol layers. Networks of quantum repeaters utilize three concepts to execute a distributed algorithm that creates entangled quantum states between nodes that are far apart: a basic entanglement mechanism which depends on the physical implementation, error management (in this work, we study a method known as purification), and finally a quantum state propagation layer (here we implement entanglement swapping, which builds multi-hop connections from single-hop connections). Some researchers are investigating approaches that are substantially different from entanglement swapping [30, 24, 19]. Here we focus on swapping, but the layered architecture approach is broadly applicable, allowing other implementations to replace only a single layer in the protocol stack. Previous work primarily focused on the physical and mathematical tools for building repeaters. Classical information is also needed to enable teleportation and swapping, as many quantum operations are not deterministic, and results of quantum measurements need to be reported to distant partners before further operations can proceed. Also, operations in the middle of the network must be coordinated to route and swap properly. This requires classical messages to make operations robust, but message propagation times penalize performance. Even though this delay is usually included in repeater simulations, prior work has not defined the protocols in detail, especially with respect to how all of the nodes make consistent decisions in a timely fashion. In this work, we introduce a protocol stack for networks of quantum repeaters that considers all the necessary classical messages and which can be easily adapted for different approaches at all three protocol layers. We run simulations of competing flows on a dumbbell topology in order to increase our confidence in the behavior of our network protocols. By adjusting the fidelity thresholds required for entanglement swapping, we show that some configurations boost the aggregate throughput for multiple flows significantly above the maximum for a single flow, taking advantage of resources that would otherwise sit idle. The operation of such complex networks and such delicate tuning of the system without formal protocol definitions would not be possible. Previous work has also concentrated almost exclusively on the dedicated use of a single line or chain of repeaters, delivering Bell pairs only to the two nodes at the ends of the chain. Networks, however, typically have more than two end points, and allow any pair of these end nodes to communicate. More topologically complex networks, with numerous end nodes, are obviously much more scalable than connecting every possible pair of nodes using a dedicated line of repeaters. we investigate how classical multiplexing schemes translate to the domain of quantum repeaters, in order to manage shared resources and active communication flows of data from different stations. We simulate four sharing protocols in a complex network with competing traffic. Circuit switching gives any individual flow the best performance, but makes poor use of the overall network and is inflexible. We use circuit switching as our baseline case to compare time division multiplexing, buffer space multiplexing, and statistical multiplexing, and show that all multiplexing schemes are better than circuit switching. For the particular network simulated, we find that statistical multiplexing outperforms time division multiplexing by 28% and buffer space multiplexing by 13%. We find that all three multiplexing schemes are fair; each flow is penalized a similar percentage of its throughput as the the total number of users in the network increases. Finally, statistical multiplexing requires no network-wide coordination of the use of quantum memory or channels, and is easier to implement than a robust, scalable scheme for the other protocols. Our current results suggest that the best strategy for quantum repeater networks is statistical multiplexing. (This is in fact the quantum analogue of the basis on which the Internet works.) However, our current simulations are done with no degradation of memory over time, using the assumption of quantum error-corrected memory at each repeater; inclusion of decoherence and a finite qubit lifetime remains as future work.
The realization of the Quantum Internet promises transformative capabilities in secure communication, distributed quantum computing, and high-precision metrology. However, transitioning from laboratory experiments to a scalable, multi-tenant network utility introduces deep orchestration challenges. Current development is often siloed within physics communities, prioritizing hardware, while the classical networking community lacks architectural models to manage fragile quantum resources. This tutorial bridges this divide by providing a network-centric view of quantum networking. We dismantle idealized assumptions in current simulators to address the "simulation-reality gap," recasting them as explicit control-plane constraints. To bridge this gap, we establish Software-Defined Quantum Networking (SDQN) as a prerequisite for scale, prioritizing a symbiotic, dual-plane architecture where classical control dictates quantum data flow. Specifically, we synthesize reference models for SDQN and the Quantum Network Operating System (QNOS) for hardware abstraction, and adapt a Quantum Network Utility Maximization (Q-NUM) framework as a unifying mathematical lens for engineers to reason about trade-offs between entanglement routing, scheduling, and fidelity. Furthermore, we analyze Distributed Quantum AI (DQAI) over imperfect networks as a case study, illustrating how physical constraints such as probabilistic stragglers and decoherence dictate application-layer viability. Ultimately, this tutorial equips network engineers with the tools required to transition quantum networking from a bespoke physics experiment into a programmable, multi-tenant global infrastructure.
The Internet has revolutionized the way we live. It has enabled applications far beyond what it was originally built for, and it will continue to exceed our expectations for the future. Quantum computers, and the network that will connect them—the Quantum Internet— are likely going to follow the same path. Differently from normal computers, quantum computers can share a property called entanglement, which allows the qubits (quantum bits) to be connected at a much more fundamental level. This property enables a range of new applications that span secure communication to enhanced metrology. Over the past fifteen years, significant progress has been made in connecting rudimentary quantum network nodes via long-distance entanglement. Several quantum platforms have demonstrated entanglement generation between two physically separated qubits. In this thesis we take a significant step forward, both in terms of experimental complexity achieved, and in the abstraction of said complexity for future developments. Moving past two-node experiments required a fundamental redesign of our experimental apparatus, as well as developing the capabilities to control simultaneously the additional node. The first result of this thesis is building a three-node entanglement-based quantum network. We demonstrated distribution of Greenberger-Horne-Zeilinger states over the network, as well as a building block for larger networks: entanglement swapping. Differently from previous multi-node demonstrations, which relied on post-selection, our network is able to perform the entanglement distribution in a heralded fashion: a signal will notify the users that the protocol was successful, and that the state is ready to be used. The second result builds on the first, by adding control over a fifth qubit, improving the quality of the entanglement, and introducing a novel repetitive readout technique, to achieve quantum teleportation of a qubit from the third node to the first—nodes that do not share a direct entanglement channel. The third and final result is the demonstration of entanglement delivery using a quantum network stack. The Internet is built using a plethora of physical platforms: optical fibers, Ethernet cables, Wi-Fi, satellite signals etc. To abstract their functionality, and make applications work regardless of the underlying platform, a layered approach was developed in the 1970s (the Internet protocol). Taking inspiration from classical network stacks, we demonstrate the first two layers of a quantum network stack, the physical layer (where the qubits, lasers and signal generators live), and the link layer, which abstracts the concepts of qubit and entanglement generation such that they can be used by applications at the higher-layers, hiding the complexity of the quantum platform being used.