⚛️ quantum physics

Capacity and SKR tradeoff in coexisting classical and CV-QKD metropolitan-reach optical links

This paper demonstrates that optimizing guardband placement for continuous-variable quantum key distribution (CV-QKD) at the edge of a metropolitan DWDM spectrum significantly improves the secret key rate by 108% while reducing capacity loss to 3.4% compared to center-band placement.

Original authors: Cagla Ozkan, Lucas Alves Zischler, Kadir Gumus, Joao dos Reis Frazao, Cristian Antonelli, Chigo Okonkwo

Published 2026-07-02

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Cagla Ozkan, Lucas Alves Zischler, Kadir Gumus, Joao dos Reis Frazao, Cristian Antonelli, Chigo Okonkwo

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

Imagine a busy highway (the fiber optic cable) where two very different types of traffic are trying to drive at the same time:

The Trucks: These are the Classical Internet signals. They are loud, powerful, and carry all your regular data (videos, emails, websites).
The Ghosts: These are the Quantum signals (specifically CV-QKD). They are incredibly fragile, carrying secret encryption keys that are mathematically unbreakable. If a single truck bumps into a ghost, the secret is ruined.

The problem is that the loud trucks create two types of "noise" that scare the ghosts:

The "Rumble" (SpRS): This is like the low-frequency vibration of heavy trucks. It happens everywhere along the road and is hard to avoid, no matter where you park.
The "Splash" (FWM): This is like water splashing from a fast-moving truck. It only happens when trucks are very close to each other and moving fast. If the trucks are too loud (high power), this splash is huge and drowns out the ghosts.

The Big Discovery

The researchers in this paper asked: "How can we let the trucks and ghosts share the same highway without the ghosts getting scared away?"

They found a clever way to arrange the traffic using Guardbands (empty lanes) and Parking Spots.

1. The "Empty Lane" Strategy (Guardbands)

Think of the highway as having 88 lanes. The researchers realized that if you put the Ghost right in the middle of the traffic, it gets splashed by trucks on both sides.

The Solution: Move the Ghost to the very edge of the highway (the last lane).
The Buffer: Leave 3 empty lanes (a "guardband") between the Ghost and the nearest trucks.

2. The Results: A Win-Win

When they tested this setup, they found a "sweet spot" for the trucks' speed (power):

If the trucks are too quiet: The "Rumble" is the main problem. The empty lanes don't help much because the vibration comes from everywhere.
If the trucks are loud (but not too loud): The "Splash" becomes the main problem. Here, the empty lanes work like magic!
- By moving the Ghost to the edge and adding 3 empty lanes, the Ghost's signal strength (SKR) doubled.
- The cost? The highway lost only 3.4% of its total space for trucks. That's like closing just a tiny parking spot to save the whole neighborhood.

3. The "One-Way" vs. "Two-Way" Traffic

They also checked if the trucks should drive in the opposite direction to avoid splashing the Ghost.

The Finding: It didn't make much difference. Driving in the same direction (co-propagation) with the empty lanes was already almost perfect. You don't need to build a separate road for the trucks to go the other way.

The Simple Rulebook

The paper concludes with a simple guide for building these networks:

If the trucks are driving slowly (low power): You don't need empty lanes. The "Rumble" is the only issue, and you can't really fix that with lanes.
If the trucks are driving fast (medium-high power): Move the Ghost to the edge of the highway and leave 3 empty lanes next to it. This doubles the security of the Ghost's message while barely shrinking the highway for the trucks.

In short: You don't need to build a new road for the internet and the quantum secrets. You just need to park the secrets at the very edge of the road and leave a few empty lanes as a buffer zone. This makes the secrets twice as strong while only taking up a tiny bit of space for the regular traffic.

Technical Summary: Capacity and SKR Trade-off in Coexisting Classical and CV-QKD Metro-Reach Links

Problem Statement
While Continuous-Variable Quantum Key Distribution (CV-QKD) offers inherent compatibility with coherent detection and higher noise tolerance compared to Discrete-Variable (DV) protocols, enabling coexistence with classical internet traffic over Dense Wavelength Division Multiplexing (DWDM) fibers, significant performance impairments remain. In metropolitan hybrid links, classical channels induce nonlinear interference that degrades the Quantum Key Rate (SKR). Specifically, Spontaneous Raman Scattering (SpRS) dominates at long distances, while Four-Wave Mixing (FWM), which scales cubically with power, becomes the dominant impairment in short-to-medium reach (10–25 km) DWDM systems. Previous studies have optimized wavelength allocation and quantified performance limits, but systematic, power-dependent guardband design rules and quantitative trade-offs between SKR and classical capacity for CV-QKD coexistence have not been thoroughly explored.

Methodology and Analytical Framework
The authors develop a comprehensive interference model for a C-band link where $N_{cl}$ classical WDM channels and one CV-QKD channel co-propagate over standard single-mode fiber. The framework utilizes the coexistence model from [7] to calculate nonlinear interference power ( $P_{int}$ ) at the quantum wavelength, accounting for SpRS, FWM, and out-of-band effects with full frequency dependence.

Key modeling components include:

Noise Mapping: $P_{int}$ is mapped to input-referred excess noise ( $\xi_{coex}$ ) for a coherent CV-QKD link using standard photon-number relations.
SKR Calculation: The asymptotic SKR is calculated for Gaussian-modulated coherent-state CV-QKD with reverse reconciliation under collective attacks, considering total noise comprising loss-induced vacuum noise, trusted detection noise, and coexistence noise.
Physical Mechanisms: The model solves coupled power-evolution equations. It highlights that FWM-induced noise scales as $\gamma^2 P^3 L_{eff}^2$ and decays with frequency separation, making it sensitive to guardband width. Conversely, SpRS scales linearly ( $g_R P L_{eff}$ ) across a broad bandwidth (~40 THz) and is largely insensitive to gigahertz-scale guardbands.
Simulation Parameters: The study sweeps guardband size ( $N_{GB}$ from 0–10 channels), per-channel launch power (−4.5 to 0.5 dBm/ch), and transmission distance (0–30 km) on an 88-channel, 50 GHz DWDM grid. The quantum channel is positioned either at the band center (Ch 44) or the band edge (Ch 88).

Key Results
The numerical analysis reveals distinct regimes of operation based on launch power and guardband configuration:

Noise Regime Transition: At low powers (e.g., −4.5 dBm/ch), SpRS dominates, resulting in an SKR of ~200 Mbit/s that remains largely independent of guardband size. At high powers (e.g., 0.5 dBm/ch), FWM dominates, driving SKR near zero without protection. A transition region exists between −5 and 0 dBm/ch where FWM overtakes SpRS.
Guardband Efficacy:
- High Power: At 0.5 dBm/ch, a 3-channel (150 GHz) guardband recovers SKR from ~0 to 38 Mbit/s.
- Intermediate Power: At −1.5 dBm/ch, a 3-channel guardband yields a 108% increase in SKR.
- Low Power: At −4.5 dBm/ch, guardbands provide negligible benefit as SpRS is the limiting factor.
Placement Strategy: Placing the quantum channel at the band edge (Ch 88) is significantly more efficient than symmetric center-band protection. For a 3-channel guardband at −1.5 dBm/ch, edge placement achieves the 108% SKR gain with only a 3.4% loss in classical capacity, compared to a 6.8% loss for center-band protection.
Propagation Direction: Counter-propagation offers only a marginal additional SKR gain (~2.6%) over co-propagation, suggesting that co-propagation with optimized guardbands is near-optimal.
Reach Extension: At high power (0.5 dBm/ch), a 3-channel guardband extends the co-propagating distance from ~3 km to ~10 km. At low power, all configurations reach 20–23 km, confirming guardbands are primarily necessary in the FWM-dominated power range.

Significance and Claims
The paper claims to provide practical, theoretically derived guidelines for deploying hybrid CV-QKD networks on existing fiber infrastructure. The primary contribution is the identification of a transition regime between FWM- and SpRS-dominated noise, which dictates the necessity and design of guardbands.

The authors assert that:

For launch powers below −2 dBm, links are SpRS-limited, and no guardband is required.
In the FWM-dominated regime (−1.5 to 0.5 dBm/ch), band-edge placement with 100–150 GHz guardbands can double the SKR while incurring only a 3.4% capacity loss.
This approach is approximately twice as efficient in terms of capacity preservation compared to symmetric band-center protection.

These findings aim to resolve the trade-off between quantum security and classical throughput, enabling scalable CV-QKD integration into metropolitan DWDM networks without the need for dedicated dark fibers.