Markdown Version | Session Recording

Session Date/Time: 03 Nov 2025 14:30

QIRG

Summary

The QIRG session at IETF 124 featured a diverse set of presentations on quantum internet research, spanning software stacks, real-time control, connectionless architectures, multi-plane frameworks, quantum datagram protocols, and foundational principles. Key themes included the challenges of moving quantum experiments from lab to field deployment, the need for standardization of network models and control, and the exploration of architectures that leverage quantum-native properties rather than merely adapting classical internet paradigms. Discussions highlighted opportunities for collaboration on open-source software and simulators, and calls for community engagement on critical research and standardization questions. Several existing drafts were discussed, and a new internet draft was introduced.

Key Discussion Points

• Administrative Items

  • Welcome to IETF 124 QIRG, first meeting on Monday morning.
  • Reminder of IRTF's focus on longer-term research issues, not standards, publishing informational or experimental documents.
  • Document Status:
    • Diego Lopez's quantum internet multi-plane proposal (individual contribution, updated version).
    • Alan Zhu and Robert Brubbergen's new draft on Quantum Datagram Control Protocol, submitted October 1st, first public discussion.

• Software Stack for Quantum Networks (Inder Monga)

  • Presented a modular software stack for quantum network control, including an in-memory database, resource orchestration, routing, and topology management.
  • Emphasized delineating orchestration (non-real-time functions on CPU) from real-time hardware control (time-critical functions on FPGAs).
  • Highlighted the need for a consistent network data model and suggested an opportunity for standardization of properties, names, node types, qubit information, and entanglement types.
  • Described a plugin architecture for routing, control, algorithms, and error correction, supported by a Hardware Abstraction Layer (HAL) for various qubit types, platforms, and simulators.
  • Vision: A framework for exploring quantum network approaches, supporting testbeds and simulation.
  • Demonstrated a core routing plugin (considering constraints like Bell states) and a two-level scheduler (automated calibration, entanglement setup).
  • Called for collaboration on existing testbeds and extending the software to implement new ideas as plugins.
  • Q&A: Discussed integration with quantum computers (work in progress) and data packaging methods (schema vs. binary, unconfirmed).

• Design and Implementation of Real-Time Control of Trapped Ion Quantum Networks (Wengee Lin)

  • Focused on trapped ion quantum networks as a leading platform due to excellent quantum information processing, long coherence times, and inherent support for entanglement distribution and storage.
  • Identified challenges for field deployment: microsecond-level qubit control, high-rate entanglement generation (hundreds of microseconds or less), precise synchronization (frequency, time, polarization, phase), and scalable/modular software.
  • Presented a two-level control framework with a network-level centralized server and node-level dedicated hardware (Arctic FPGAs) for hard real-time control.
  • Software is modular, supporting repeatable execution, scans, and calibration for quantum operations (e.g., Rabi, Ramsey oscillations).
  • Posed questions to the community: What are the major challenges for field deployment? What basic building blocks need standardization (architecture, control)? Where should the demarcation line between non-real-time and real-time control be drawn in quantum networks, similar to TCP/IP?
  • Q&A: Clarified that distributed synchronization for clocks and high-level control is managed through dedicated links and control systems (e.g., urea cavities for frequency synchronization), not solely by node-level FPGAs.

• InterQNet: Heterogeneous Full-Stack Approach to Co-Designing Scalable Quantum Networks (Joaquin Chung)

  • Introduced the InterQNet project, comprising "InterQNet Achieve" (campus-scale testbed at Argonne National Lab with Yb ions, Er ions, superconducting qubits, and automated frequency conversion) and "InterQNet Scale" (system study using simulations and theory for scalable architectures).
  • InterQNet Scale work: Presented an Adaptive Continuous Entanglement Generation Protocol (ACP) for continuously generating entanglement between neighbors, adaptively focusing on popular user requests. Simulations on "Sequence" (a discrete event simulator) showed ACP significantly improved "time to serve" and maintained fidelity compared to on-demand and uniform continuous protocols.
  • InterQNet Achieve work: Demonstrated remote entanglement distribution across a campus-scale quantum network testbed, utilizing a multi-plane abstraction (infrastructure, control, service planes). The control plane includes agents for devices, communicating via GRPC with a centralized orchestrator for automated calibration and continuous operation.
  • Achieved 95% fringe visibility over three remote sites and successfully demonstrated automated polarization compensation.
  • Q&A: Discussed the choice of simulation parameters (e.g., optimistic 2-second decoherence time) and its potential impact on protocol performance comparison.

• Connectionless Quantum Networks (Leonardo Bachetetti)

  • Proposed a connectionless architecture for two-way quantum networks to simplify operations, reduce synchronization overhead, and achieve high utilization, leveraging lessons from the classical internet.
  • Defined "connectionless" as an application where state is not kept at intermediate nodes, requests are treated separately, and there's no long-term resource reservation.
  • Proposed sequential swaps schedules for realizing connectionless operation, abstracting the movement of quantum and classical information as a "quantum datagram" (QDatagram) – entanglement swapping becomes QDatagram forwarding.
  • Drew an analogy to packet-switched systems, suggesting that managing QDatagram congestion can control decoherence.
  • Positioned this within a quantum internet protocol stack, with a transport layer managing decoherence via a Quantum TCP (QTCP)-like protocol (AIMD congestion control based on QDatagrams), showing it outperforms quantum memory reservation schemes.
  • Introduced Q Primal Dual, a quantum-aware protocol that uses entanglement generation rate/quality trade-offs as a control knob, demonstrating significant improvements in secret key rate.
  • Highlighted "Quantum Savory," an open-source simulator developed at UMass Amherst.
  • Q&A: Clarified that the classical information for BSM operations can use TCP/IP or faster protocols, and the model is agnostic to the classical transport, with quantum and classical components being abstractly mapped rather than physically coupled.

• Leveraging Internet Principles for Quantum Networks (Marcello Galefi)

  • Argued that RFC 9340's abstract goal of building a quantum network stack from the ground up provides an opportunity to challenge classical internet principles (simplicity, end-to-end argument) which are ill-suited for the stateful nature of entanglement.
  • Advocated for a radical departure from TCP/IP, moving from packet switching to entanglement switching, where the core goal is to entangle nodes and distribute correlations, not forward information packets.
  • Proposed bringing "intelligence within the network" (circuit switching analogy) due to entanglement's statefulness.
  • Introduced concepts of an Entanglement Service Provider (ESP) and an Entanglement Defined Control plane, emphasizing a quantum-native, potentially centralized control plane.
  • Criticized existing quantum control plane proposals for following classical implicit assumptions, leading to scalability and suboptimality issues, and for not exploiting entanglement's ability to redefine "neighborhood."
  • Proposed a quantum native control plane with quantum network addresses and a quantum header/payload for implementing control and distributing entanglement.
  • Presented a case for quantum addressing in achieving the first fully name-independent compact routing for quantum networks (achieving an entangling stretch of 3-5).
  • Issued a call for interest to the community to work on quantum control plane concepts, noting that RFC 9340 explicitly declared this out of scope, and expressed openness to collaboration.

• Quantum Tomography, Teleportation, and Distributed Quantum Computing (Roger Selly)

  • Presented a vision for scalable distributed quantum computing by linking many smaller quantum computers.
  • Introduced the Quantum Possibility Space (QPS), a compact and accurate representation of quantum tomography, defining wave functions of multi-qubit entangled states using non-negative integers (state vectors) and complex magnitudes. Demonstrated JSON encoding.
  • Used Gilles Brassard's quantum teleportation as a hands-on example, visualizing state probability changes gate-by-gate through Alice's and Bob's circuits.
  • Analyzed the functional takeaway from the teleportation circuit, describing how Alice's qubit 0 is teleported to Bob's qubit 2, conditioned on the state of Alice's qubit 2.
  • Demonstrated web-based quantum teleportation, sending QPS data via JSON Web Signature/Token over the conventional internet.
  • Discussed application of QPS for Quantum Key Distribution (QKD) using its natural conjugate property to confirm entanglement.
  • Explained how QPS evolves upon measurement and collapse from a superposition state.

• Multi-Plane Architecture Proposal (Diego Lopez)

  • Presented an updated iteration of the multi-plane architecture framework draft.
  • Goal: Provide a general framework for reasoning about protocol definitions, APIs, classical/quantum alignment, and coordination, leveraging operational experience from QKD deployments.
  • Architecture: Based on SDN concepts, it is multi-plane and multi-strata, dealing with physical connectivity, classical networks, and pure quantum effects.
  • Introduced the concept of a "Service Unit" as a neutral term to identify a shared state between two application entities (e.g., bit strings for QKD, entangled pairs for general entanglement), analogous to transport flows and identified by UUIDs.
  • Updates in the draft:
    • Identified protocols and interfaces for physical connectivity (optical, ACTN, Path Computation Elements) and quantum forwarding (e.g., quantum-native addressing, quantum repeaters).
    • Elaborated on how service units map physical and entanglement overlays and how their parameterization is managed.
    • Introduced the concept of synthetic environments (emulation) to analyze proposals beyond the limited scale of current testbeds and simulations. Aims to emulate distributed quantum systems and their interaction with classical mechanisms, potentially at a scale of thousands of nodes, and capable of mixing real and emulated nodes.
  • Call for adoption: Believes the draft is mature for adoption as a research group document, not as a standard, but as a framework and common ground for discussing and evaluating protocols and identifying research issues.
  • Chair Comment: The current draft contains significant "state of the art" and "thinking out loud" content which needs refinement for long-term utility as a framework document.

• Quantum Datagram Control Protocol (QDCP) (Alan Zhu)

  • Introduced a new internet draft for a Quantum Datagram Control Protocol, building on successful experiments in hybrid classical+quantum internet routing.
  • Datagram Procedure: Designed for the no-cloning theorem, where quantum information is a single data block. It consists of a lightweight classical control metadata packet followed by the quantum packet.
  • Packet Design: Built on UDP, it includes a UDP header, a QDCP header (version, flags, length, reserved bits), followed by Type-Length-Value (TLV) payloads. Nine TLV types are currently defined, and they can be chained.
  • Key TLV Examples:
    • 04: Reconfigurable Optical Add-Drop Multiplexer (ROADM) output port ID (for switch configuration).
    • 05: Quantum packet delay (time between classical and quantum packets).
    • 06: Duration of quantum information (for maximizing signal-to-noise ratio).
    • 09: Polarization correction (for compensating fiber-induced polarization shifts using classical light).
  • Demonstrated a TLV chaining example for configuring a ROADM, specifying quantum information duration, packet delay, and polarization correction, followed by the quantum information itself. The quantum information length is not limited by the classical header.
  • Security Concerns: QDCP inherits UDP risks (spoofing, injection, replay) for classical information, necessitating trusted environments or DTLS/IPSEC protection. TLVs can reveal network state. Future work on TCP/Quick compatibility is being considered.
  • Chair Comment: A new -01 version of the draft will be submitted and associated with the initial -00 version.

Decisions and Action Items

• QIRG Co-Chairs:
  • To follow up on the potential adoption of Diego Lopez's multi-plane architecture draft.
  • To assist Alan Zhu in associating the new -01 version of the QDCP draft with the previous submission in the Datatracker.
  • Plan to hold a QIRG meeting at IETF 125 in Shenzhen.
• Inder Monga: Continue discussions on integrating the software stack with quantum computers.
• Diego Lopez: Refine the multi-plane architecture draft to focus on its role as a framework for long-term use, potentially reducing "state of the art" and "thinking out loud" content.
• Alan Zhu: Submit the -01 version of the Quantum Datagram Control Protocol draft.

Next Steps

• Community Engagement:
  • Discuss the need for standardization of a consistent network data model for quantum networks.
  • Address Wengee Lin's questions on major challenges for field deployment, basic building blocks for standardization, and the demarcation line between real-time and non-real-time control.
  • Engage with Marcello Galefi's call for interest in a quantum-native control plane, particularly in light of RFC 9340's current scope.
  • Explore collaboration opportunities on Inder Monga's open-source software stack and Leonardo Bachetetti's Quantum Savory simulator.
• Draft Progression:
  • Further discussion and potential adoption of the multi-plane architecture draft as a QIRG document.
  • Continue development and community review of the Quantum Datagram Control Protocol draft.
• Future Meetings:
  • Look forward to the QIRG meeting at IETF 125 in Shenzhen.