**Session Date/Time:** 07 Sep 2022 12:00

# [QIRG](../wg/qirg.html)

## Summary
This session featured a presentation by Stephen Diadamo on the concept of packet switching in quantum networks as a pathway toward a scalable quantum internet. The presentation covered existing quantum network architectures (QKD-specific and entanglement-based), highlighted their limitations, and proposed a vision for universal, transparent, and scalable packet-switched quantum networks, including an intermediate burst-switching approach and a hybrid network model. The discussion included technical challenges and simulation results for key quantum applications.

## Key Discussion Points

*   **Introduction to Quantum Networks:**
    *   A quantum network connects quantum-capable devices, possibly alongside classical nodes.
    *   Channels typically use free space or fiber optics, with photons as the transport medium.
    *   An OSI-like layered model is commonly proposed for quantum network stacks.
    *   Networks can operate at local (data center), metropolitan (city-wide QKD), or wide-area (interconnecting quantum networks with quantum repeaters or satellites) scales.
*   **Applications of Quantum Networks:**
    *   **Quantum Key Distribution (QKD):** Near-term, well-established application for secure key exchange.
    *   **Blind Quantum Computing:** Performing quantum computations on a remote quantum computer without revealing the algorithm.
    *   **Distributed Quantum Computing:** Connecting quantum computers to perform larger-scale algorithms.
    *   **Networked Quantum Sensors:** Enhancing sensitivity for measuring time, magnetic fields, gravity, etc., by connecting quantum sensors.
*   **Speaker's Group Focus (Cisco Quantum Research Group):**
    *   Focused on lower layers of the network stack: protocols for quantum state transport, quantum network simulation, control, security, and hardware (optical routers/switches, quantum repeaters, quantum memory).
*   **Existing Quantum Network Architectures:**
    *   **Quantum Key Distribution (QKD) Networks:**
        *   Application-specific (only for QKD), often deployed side-by-side with classical networks.
        *   Utilizes "trusted repeaters" (trusted relays) to overcome distance limitations due to lack of true quantum repeaters.
    *   **Entanglement-Based Quantum Networks:**
        *   Relies on establishing entanglement between nodes, then using quantum teleportation to transport quantum states.
        *   General purpose (can send any quantum state), but technologically challenging (robust entanglement distribution and storage, potentially slow rates).
        *   Often envisioned as working alongside the classical internet.
*   **Limitations of Existing Designs:**
    *   **Scalability:** Rely on circuit switching or static switching models, which historically don't scale well for many users.
    *   **Application Specificity:** QKD networks are often single-purpose; entanglement-based might not be optimal for prepare-and-measure QKD.
    *   **Integrability:** Limited integration with existing classical internet infrastructure and protocols.
*   **Vision for Future Quantum Networks: Packet Switching**
    *   **Three Pillars:** Universality (support all quantum applications), Transparency (integrate with classical technology, e.g., shared fibers/hardware), Scalability (scalable protocols and state transport).
    *   **Packet Switching Concept:**
        *   Contrast with circuit switching (dynamic vs. reserved routes, ordered vs. out-of-order arrival, fair use).
        *   Proposed **Hybrid Data Frame:** Classical Header + Quantum Payload + Classical Trailer.
        *   Allows dynamic routing of quantum information.
    *   **Multiplexing Schemes:** Time Division Multiplexing (TDM) and Wavelength Division Multiplexing (WDM) for header, payload, and trailer.
    *   **Node Architecture:** Classical control unit triggers C-Q transmissions. At intermediate nodes, demultiplexing occurs, the classical header is processed, the quantum payload is stored in a quantum memory or delay line, and then recombined and re-multiplexed.
    *   **Long-Term Vision:** Quantum Reconfigurable Optical Add/Drop Multiplexers (ROADM) incorporating quantum memories, error correction, and frequency conversion.
*   **Challenges for Packet Switching in Quantum Networks:**
    *   **No Quantum Amplification:** Requires extremely low-loss optical hardware.
    *   **Distance Limitations:** Switches may need to be closely spaced (e.g., single-digit kilometers).
    *   **Crosstalk:** Minimizing Raman noise between strong classical and weak quantum signals sharing a fiber.
    *   **Hardware Constraints:** Integration of classical and quantum switching in devices, potentially requiring cryogenic temperatures.
    *   **Networking Specifics:** Robust quantum memories, quantum error correction, choice between quantum relays (dumb forwarders) and repeaters (complex processing), network security.
*   **Intermediate Compromise: Burst Switching:**
    *   Sends classical header ahead of time to prepare the switch, minimizing the need for quantum memory storage at intermediate nodes for the quantum payload.
    *   Involves precise classical-quantum transmission scheduling and channel reservation (less stringent than circuit switching, more than pure packet switching).
    *   Can mitigate crosstalk effects.
    *   Challenges include determining optimal burst time (guard time) based on network parameters and number of hops.
*   **Hybrid Network Model:**
    *   Combine packet switching for short-distance segments with entanglement-based "teleportation channels" (first-generation quantum repeaters) for longer distances.
*   **Simulation Results:**
    *   **QKD (BB84) over a line network:** Simple model shows positive secret key rates are achievable over multiple hops, even with some loss (optimistic scenario).
    *   **Entanglement Distribution (Bell states):** NetSquid simulations show achievable fidelity with noisy quantum memories and channels, particularly at shorter distances. Millisecond-regime memory storage appears necessary. Processing time for packets at nodes also critical.
*   **Open Problems and Future Outlook:**
    *   Crosstalk mitigation (especially Raman noise) in network and protocol design.
    *   Identifying scenarios where packet switching outperforms other designs and defining relevant metrics.
    *   Efficient quantum frame generation and guard time optimization for burst switching.
    *   Reducing classical header processing time and defining essential header information.
    *   Emphasized the importance of considering user scalability and hybrid approaches now, based on lessons from classical internet history favoring packet switching.

*   **Q&A Discussion:**
    *   **Classical Header/Quantum Payload Separation and Loss:** Separated by time or wavelength multiplexing. Loss management depends on the application; quantum payloads lost if header is lost. Retransmission may be possible for QKD, but complex for quantum computing.
    *   **Quantum Error Correction (QEC):** QEC involves encoding logical qubits onto multiple photons (a "set of photons") that would travel together as part of the quantum payload. This is a general challenge for quantum repeaters, not specific to packet switching.
    *   **Routing Schemes:** The proposed packet-switching model can accommodate various dynamic routing approaches, including source routing or label switching, not just hop-by-hop.
    *   **Wavelengths for Classical/Quantum Coexistence:** Discussion on using 1310nm for quantum and 1550nm for classical traffic to reduce crosstalk, even with good multiplexing. Hollow-core fiber was mentioned as a potential solution for mixing signals with less crosstalk, though it currently has higher loss.
    *   **Standardized Icons:** A request was made for Stephen Diadamo to consider adopting a set of standardized icons for quantum network node types for consistency across research groups.

## Decisions and Action Items

*   Stephen Diadamo was requested to share a link to the proposed standardized quantum network icons for consideration.
*   Discussions on the presented material are encouraged to continue on the QIRG mailing list.

## Next Steps

*   The QIRG plans to continue hosting seminars approximately twice a year, with announcements posted on the mailing list.
*   Participants were encouraged to attend the upcoming IETF meeting in Tokyo in March.