Multiverse: A Simulator for Evaluating Entanglement Routing in Quantum Networks
Amar Abane, Junxiao Shi, Van Sy Mai, Abderrahim Amlou, Abdella Battou · 2025
Read Multiverse: A Simulator for Evaluating Entanglement Routing in Quantum Networks by Amar Abane, Junxiao Shi, Van Sy Mai, Abderrahim Amlou, Abdella Battou (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-22937.txt.
Introduction
Amar Abane, Junxiao Shi, Van Sy Mai, Abderrahim Amlou, and Abdella Battou present MQNS, a discrete-event simulator for testing how entanglement should be routed through quantum networks. Routing choices interact with purification, swapping, and memory management. MQNS lets researchers configure all of them at run time inside one qubit lifecycle model, with integrated link-architecture models. Its modular, minimal design stays independent of any single network architecture. Competing routing paradigms can therefore be compared fairly and reproducibly on the same footing.
Catalog record
| CREATOR | Amar Abane, Junxiao Shi, Van Sy Mai, Abderrahim Amlou, Abdella Battou |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-22937
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multiverse: A Simulator for Evaluating Entanglement Routing in Quantum Networks Amar Abane, Junxiao Shi, Van Sy Mai, Abderrahim Amlou, Abdella Battou Abstract We present MQNS, a discrete-event simulator for rapid evaluation of entanglement routing under dynamic, heterogeneous configurations. MQNS supports runtime-configurable purification, swapping, memory management, and routing, within a unified qubit lifecycle and integrated link-architecture models. A modular, minimal design keeps MQNS architecture-agnostic, enabling fair, reproducible comparisons across paradigms and facilitating future emulation. I Introduction Quantum networks (QNs) enable distributed computing and quantum key distribution by generating fragile long-distance entanglement despite photon loss, decoherence, and noisy operations. End-to-end distribution combines routing, swapping, and purification across heterogeneous hardware (color centers, ions, superconductors). Designing networks that account for diverse hardware processes, encoding schemes, and interdependent operations remains a major challenge. While different entanglement routing paradigms have been proposed [ 1 ] their evaluation remains limited by the lack of comprehensive, configurable simulation tools. Although several simulators exist, most entanglement routing studies rely on custom single-purpose simulators or analytical models that fail to capture dynamic behaviors or the interplay of repeater functions. As QNs are increasingly viewed as communication systems—drawing on classical networking concepts—simulators must support comparative evaluations of network architectures and routing paradigms. We introduce Multiverse Quantum Network Simulator (MQNS) 1 1 1 https://github.com/usnistgov/mqns , a discrete-event simulator that enables quick evaluation of entanglement distribution under dynamic configurations. MQNS offers configurable routing algorithms, swapping strategies, purification schemes, and memory management policies. It models a large set of physical link architectures and encoding schemes for Einstein–Podolsky–Rosen (EPR) pairs generation. The remaining of this article reviews existing simulators and their limitations, introduces the core functions and key conceptual components of MQNS. Implementation aspects and a demonstration architecture are then described, followed by validation results and use cases. II Related Work SimulaQron [ 10 ] provides an emulation-like environment for application-layer testing but lacks built-in routing functionalities. While NetSquid [ 9 ] offers quantum state tracking and interfaces for protocol development, it requires extensive customization for advanced routing. SeQUeNCe [ 17 ] is recognized for its precision and modularity but is limited to static swapping and routing configurations. QuNetSim [ 12 ] is designed for small-scale QNs with an OSI-inspired architecture abstracting the network behavior. Consequently, QuNetSim is best suited for evaluating upper-layer quantum applications rather than routing and architectures. QuISP [ 15 ] provides scalable modeling with error tracking and introduces the RuleSet protocol for programmable entanglement distribution. However, its reliance on the RuleSet architecture makes dynamic routing dependent on protocol extensions. SimQN [ 7 ] adopts an architecture-agnostic design, aligning with the current lack of a recognized QN architecture and protocol stack. Its limitation lies in a design that does not accommodate inter-paradigm evaluation and a rigid implementation of critical functions such as swapping and purification. Each simulator reflects evolving research needs, progressing from application-oriented tools to network-level solutions inspired by classical networking. While some simulators can be extended to support functions such as routing, swapping, and purification, their configurations are typically static or tied to specific protocol designs. MQNS differs by unifying these functions into a single framework with runtime reconfigurability, supporting diverse routing paradigms and network architectures within the same platform. III Design Approach MQNS targets fast evaluation of QN architectures and network-layer mechanisms. Its minimal design keeps it architecture-agnostic. By avoiding artificial constructs such as global resource managers, it allows realistic architecture and protocol design and can evolve into a lightweight emulation platform. MQNS design is based on the following functions. Entanglement routing strategies follow two main paradigms: proactive and reactive, both supporting distributed or centralized deployment [ 1 ] . In proactive routing, paths are computed and installed before the forwarding phase, which includes EPR generation, purification, and swapping. In reactive routing, nodes continuously attempt elementary EPR generation on all physical channels. Routing is performed on the instant topology formed by generated EPRs, then informs the forwarding operations. Swapping strategies, in the form of static orders such as sequential or balanced-tree and policies such as swap-as-soon-as-possible (swap-asap) should be configurable at runtime through routing instructions. Purification schemes can also be dynamically configurable, allowing the selection of elementary links or segments—EPRs that span two or more elementary links—to purify for a certain number of rounds. Memory management policies determine how qubits are assigned to physical channels at topology creation. Multiplexing schemes define how qubits are allocated to paths and which qubits are selected for operations when multiple candidates are available. IV Key Components MQNS is structured around four conceptual components that can be customized and composed to realize diverse networking scenarios. IV-A Time Scheduling Entanglement distribution can be coordinated in an asynchronous or synchronous (slotted) timing mode. As illustrated in Figure 1 , asynchronous EPR generation can start as soon as two neighboring nodes have a pair of available qubits. Swapping and purification can take place anytime relevant EPRs are available, in accordance with the routing instructions. In synchronous mode, each time slot T_{s} is divided into three phases: an external phase T_{ext} for EPR generation, an internal phase T_{int} for entanglement swapping and purification, and an application phase T_{app} for users to consume E2E entanglements. In reactive routing under synchronous mode, a dedicated routing interval T_{r} is introduced after T_{ext} to collect the instant entanglement topology and retrieve routing instructions (e.g., from a controller) before the internal phase begins. Fig. 1: Timing modes in entanglement routing. Synchronous (left) and Asynchronous (right) IV-B Memory Management Memory management supports entanglement distribution needs from static and predefined to dynamic and adaptive configurations. Qubits can be assigned to channels at topology creation, either in a symmetric way—assigning the same number of qubits to both ends of a channel—or asymmetric. A two-way reservation exchange is performed between each pair of neighbors prior to EPR generation. This process is initiated by the primary node (defined as the left side of the channel) once a free qubit becomes available, typically after a qubit is consumed or decoheres, and completed when the secondary node confirms availability on its end. This process is immediately followed by the EPR generation. At runtime, each node uses routing instructions to pre-allocate qubits to installed paths either exclusively (blocking multiplexing) or by splitting qubits across multiple paths (buffer-space multiplexing). Alternatively, paths may be installed without predefined qubit allocations, allowing nodes to manage their local qubits autonomously. In such cases, statistical multiplexing assigns a newly generated EPR to the first available EPR waiting for swapping. IV-C Qubit Lifecycle Qubits follow a finite state machine [ 3 ] depicted in Figure 2 . Each qubit begins in the RAW state and transitions to the ACTIVE state during the qubit reservation process. When two qubits are reserved for EPR generation, they enter the RESERVED state. Once an EPR is generated, both qubits move to the ENTANGLED state. A qubit in the PURIF state means that its EPR satisfies swapping conditions but has to apply purification configuration first. If purification is required, the qubit undergoes the number of specified rounds looping between PURIF and PENDING states. When the purification requirements are satisfied, the qubit advances to the ELIGIBLE state. If purification fails during the process, the qubits are released. Purification is coordinated using classical message exchange.