NetFI - Gilbert-Elliott Network Packet Loss Emulator

A Python implementation of Gilbert-Elliott packet loss models with advanced network emulation capabilities.

Key Features

Gilbert-Elliott Models

• Classic 2-State GE Model - Basic Markov chain model with Good/Bad states
• 3-Parameter Extended GE - Enhanced model with additional transition parameters
• 4-State HMM Model - Advanced Hidden Markov Model with Pareto-distributed Burst Length Loss (BLL)

Advanced Packet Loss Modeling

• Pareto BLL Model - Implements heavy-tailed Pareto distribution for burst loss characterization
• Cross-Layer Emulation - Considers both physical layer (SNR) and MAC layer (channel occupation) parameters
• Real-time Visualization - Integrated matplotlib visualizations for loss patterns and burst statistics

Network Emulation

• Programmatic Packet Loss Injection - UDP/TCP packet manipulation using Scapy
• GUI Configuration - Real-time parameter adjustment through dedicated interface
• Logging & Analysis - LZ4-compressed packet logs with statistical analysis tools

Installation

System Requirements:

• Linux kernel 4.4+ (for raw socket access)
• libpcap-dev (Scapy dependency)
• Tested with Python 3.8-3.10

Usage

Launch GUI Configuration:

python src/netfi/configurationUI.py

GUI Configuration Interface:

Launch Network Emulator:

python src/netfi/network_emulator.py --input-port 5000 --output-port 5001 --protocol udp --model GE_Pareto_BLL --config config.json --log packet_logs.bin

Programmatic Emulator Control:

from netfi.emulators.packet_loss_emulator import PacketLossEmulator


emulator = PacketLossEmulator(
    input_port=6000,
    output_port=6001,
    output_ip='127.0.0.1',  # Destination IP address
    model_name='GE_Pareto_BLL',
    params_path='config.json',
    log_packets=True,  # Enable packet logging
    log_path='packet_log.bin'  # Path to save packet logs
)
emulator.start()

Configuration File Example (config.json):

{
    "GE_Pareto_BLL": {
        "Good": {
            "transitions": {
                "Good": 0.75,
                "Bad": 0.05,
                "Intermediate1": 0.1,
                "Intermediate2": 0.1
            },
            "distribution": "pareto",
            "params": {
                "alpha": 3.0,
                "lambda": 6.0
            }
        },
        "Bad": {
            "transitions": {
                "Good": 0.05,
                "Bad": 0.75,
                "Intermediate1": 0.1,
                "Intermediate2": 0.1
            },
            "distribution": "pareto", 
            "params": {
                "alpha": 3.0,
                "lambda": 6.0
            }
        },
        "Intermediate1": {
            "transitions": {
                "Good": 0.1,
                "Bad": 0.1,
                "Intermediate1": 0.7,
                "Intermediate2": 0.1
            },
            "distribution": "pareto",
            "params": {
                "alpha": 3.0,
                "lambda": 6.0
            }
        },
        "Intermediate2": {
            "transitions": {
                "Good": 0.1,
                "Bad": 0.1,
                "Intermediate1": 0.1,
                "Intermediate2": 0.7
            },
            "distribution": "pareto",
            "params": {
                "alpha": 3.0,
                "lambda": 6.0
            }
        }
    }
}

Documentation

Model Implementations

1. Classic Gilbert-Elliott (GE)

   • 2-state Markov model (Good/Bad states)
   • Parameters: p (G→B), q (B→G)

2. GE-Pareto BLL Hybrid

   • 4-state Hidden Markov Model
   • Pareto Type II distribution for burst lengths
   • Parameters: shape (α), scale (λ)

3. Network Emulator

   • Real-time packet manipulation
   • Supports both UDP and TCP protocols
   • Configurable through JSON or GUI

References

1. Bacioccola, A., Cicconetti, C., & Eklund, C. (2010). "The Gilbert-Elliott Model for Packet Loss in Real Time Services on the Internet." IEEE Xplore

2. da Silva, C.A.G., & Pedroso, C.M. (2022). "Packet Loss Characterization Using Cross Layer Information and HMM for Wi-Fi Networks." MDPI Sensors

3. Hammarhjelm, E., & Jönsson, A. (2023). "AI-Based Self-Adaptive Software in a 5G Simulation." DiVA

4. Pieper, J., & Voran, S. (2023). "Relationships between Gilbert-Elliot Burst Error Model Parameters and Error Statistics." NTIA Technical Memorandum TM-23-565

Contributing

Contributions are welcome! Please follow:

1. Fork the repository
2. Create your feature branch
3. Commit changes with descriptive messages
4. Push to the branch
5. Submit a pull request

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Project Maintainer: Hamid Reza Roodabeh Email: hr.roodabeh@gmail.com
Project Link: https://github.com/hamidrezaro/NetFI