This repository provides implementations and simulations of BB84, E91, and MDI-QKD quantum key distribution protocols using the NetSquid discrete-event simulator.
It is designed as a research-oriented toolset to simulate realistic quantum communication networks, focusing on physical-layer imperfections, fiber-channel noise, and measurement device limitations.
During my final year research project, I identified a major challenge in QKD research:
There were no publicly available, accurate, and realistic QKD simulation codes.
Most open-source QKD simulators:
- Ignore realistic channel noise and photon loss
- Simplify QKD logic without true physical modeling
- Lack modularity for research-level extension
Implementing QKD experimentally requires expensive hardware:
- Single-photon sources, detectors, and entanglement devices
- Precision optical fiber setups
- Sophisticated synchronization and error correction systems
This project was built to bridge that gap by providing:
- Physically realistic QKD simulations
- Customizable models of quantum devices and channels
- Research-ready modular code for extending QKD protocols
| Set | Description |
|---|---|
| Advanced Set | Modular implementation with device-level customization and detailed noise models. Ideal for research experiments. |
| Basic Set | Lightweight, single-file implementations for quick demonstrations and testing. Beginner-friendly. |
Both sets support performance visualization with plots (e.g., QBER, throughput) for a 10 km optical link.
- Simulator: NetSquid
- NetSquid Version Used:
1.1.1 - Programming Language: Python 3.x
- Core Dependencies:
netsquidnumpymatplotlib
| Feature | NetSquid | QuNetSim | SimulaQron | Qiskit/Cirq |
|---|---|---|---|---|
| Simulation Approach | Discrete-event simulation with detailed physical modeling | High-level Python interface for protocol design | Application-focused, limited low-level modeling | Quantum circuit paradigm (gate/circuit simulation) |
| Physical Layer Realism | High: Supports hardware imperfections, photon loss, qubit decoherence | Low: High abstraction, not physics-focused | Low: Minimal physics modeling | Low: Focused on circuit logic, not networks |
| Scalability | High: Designed for large network simulations | Moderate: Suitable for small-scale networks | Limited: Application development focus | Limited: Focused on individual circuits |
| Key Advantage for QKD Research | Essential for simulating hardware noise & channel imperfections | Good for simple network logic | Minimal support for physical realism | Not designed for network-level QKD simulations |
Why NetSquid? I selected NetSquid because of its high physical-layer realism, noise modeling, and scalability, making it ideal for accurately simulating QKD protocols and quantum network imperfections.
| Protocol | Description | Reference |
|---|---|---|
| BB84 | The first QKD protocol (Bennett & Brassard, 1984), using photon polarization states to generate secure keys. | Comprehensive Study of BB84 |
| E91 | Uses entangled photon pairs and Bell inequality tests for security verification, resistant to eavesdropping. | Wikipedia - E91 Protocol |
| MDI-QKD | Measurement-Device-Independent QKD eliminates detector side-channel attacks via a Bell-state measurement (BSM) node. | arXiv:1109.1473 |
The simulations are configured for a 10 km optical fiber channel between Alice and Bob, with the following roles:
- Alice: Prepares and transmits qubits using a quantum processor.
- Bob: Measures incoming qubits using a quantum processor.
- Charlie (MDI-QKD): Performs Bell-State Measurement (BSM) to detect entanglement using a CNOT + Hadamard measurement sequence.
- Entanglement Source: Generates entangled photon pairs for protocols like E91.
- Bell Parameter: Derived from CHSH Bell inequality tests to verify non-classical correlations.
This project includes advanced noise and device models to simulate real-world constraints:
| Model | Description |
|---|---|
| Fiber Delay Model | Simulates finite speed of light in optical fiber (c = 2 × 10^5 km/s). |
| Fiber Loss Model | Implements probabilistic qubit loss (p_loss_length = 0.2, meaning 0.2 dB/km attenuation). |
| Depolarization Noise Model | Applies qubit depolarization noise (depolar_rate = 0.01). |
| Classical Communication | Simulates basis exchange messages and tracks communication overhead. |
| Random Basis Generator | Creates random basis sequences for Alice and Bob (0/1 bit encoding). |
| Bell-State Measurement (BSM) | Charlie node detects entangled states via CNOT-Hadamard measurement sequences. |
These features allow realistic performance evaluation of QKD protocols in fiber networks.
| Metric | Definition |
|---|---|
| Throughput (qubits/sec) | Number of successfully transmitted qubits per second. |
| Synchronization Time (ms) | Time required to synchronize Alice and Bob’s measurement bases. |
| Raw Key Rate (qubits) | Amount of raw key material before error correction. |
| QBER (%) | Quantum Bit Error Rate; lower values indicate better channel quality. |
| Latency (ms) | Total delay in qubit transmission and key generation. |
| Communication Overhead | Number of classical messages exchanged for authentication and error correction. |
| Channel Loss Rate (%) | Percentage of qubits lost during transmission. |
Example -
This project is licensed under the MIT License - see the LICENSE file for details.