aqnet — Adversarial Queue Networks

Mathematical Modeling and Performance Analysis of Multi-Node Queueing Networks Under Adversarial Attacks and Timeout-Driven Retransmissions
Author: Ilias Chrysovergis (Imperial College London & Metatopia)
Preprint / Target: IEEE Transactions on Networking (ToN) / arXiv (cs.NI, math.PR, cs.CR)

---

⚡ Quickstart

Installation

Install via pip:

pip install aqnet

Or install in development mode from source:

git clone https://github.com/iliachry/Attacks-Delay.git
cd Attacks-Delay
pip install -e .

Python API Usage

import aqnet

# 1. Closed-form analytical delay calculation (Case 1: Pre-service Destruction)
theory_delay = aqnet.solve_one_node_destruction(
    lambda_n=2.0,  # Fresh packet arrival rate
    p=0.2,         # Attack probability
    mu=10.0,       # Server capacity
    T=2.0          # Retransmission timeout window
)
print(f"Expected Sojourn Time: {theory_delay:.4f} s")

# 2. Run SimPy discrete-event simulation
sim_delay = aqnet.simulate_one_node_destruction(
    lambda_n=2.0,
    p=0.2,
    mu=10.0,
    T=2.0,
    sim_duration=5000.0,
    seed=42
)
print(f"Simulated Sojourn Time: {sim_delay:.4f} s")

# 3. Analyze multi-hop tandem chain (Case 3)
tandem_delay = aqnet.solve_tandem_theory(p=0.1, N=3, mu=2.0, lambda_arrival=0.15, W=8.0)
print(f"Tandem (N=3) Delay: {tandem_delay:.4f} s")

Command Line Interface (CLI)

aqnet comes with a built-in CLI for rapid experimentation and benchmarking:

# Run benchmark across all 5 topologies:
aqnet bench

# Simulate specific topology:
aqnet run --topology tandem --nodes 3 --p 0.1 --lambda-arr 0.15 --reps 20

---

Overview

The reliable transport of data through adversarial networks requires a rigorous mathematical understanding of the coupled dynamics between malicious disruption and transport-layer recovery protocols.

This repository and package provides the complete analytical framework, discrete-event simulation engine (SimPy), benchmark dataset, and LaTeX manuscript files for studying packet sojourn times across five foundational network topologies subject to active packet destruction and modification attacks.

By synthesizing classical queueing theory, renewal-reward theory, and fixed-point traffic conservation, this project validates closed-form and semi-analytical models against discrete-event simulations, demonstrating exact agreement across all operational regimes with relative errors consistently below 1%.

---

Key Topologies & Case Studies

Topology / Case Study	Attack Model	Analytical Tool	Key Dynamic	Script Location
Case 1: 1-Node Destruction	Pre-service packet destruction	Renewal-Reward & Tail Probabilities	Packet destroyed before service; timeout/backoff trigger retransmissions	1_one_node_destruction/one_node_packet_attack.py
Case 2: 1-Node Modification	Post-service payload corruption	Fixed-Point Traffic Conservation	Packet completes service; corruption detected at receiver; full server cycle wasted per attempt	2_one_node_modification/one_node_packet_modification.py
Case 3: Tandem Chain	Multi-hop stage-wise attacks	Hypoexponential Sojourn Matching	Multi-server journey times with cumulative attack probabilities	3_tandem_chain/tandem.py
Case 4: Feedforward Network	$N$-node pipeline attacks	Gamma Moment Matching & Truncated Expectations	Non-linear traffic amplification loops, stage-wise attenuation, and stability envelope contraction	4_n_node_feedforward/n_node_feedforward.py
Case 5: Feedback Mesh	$N$-node symmetric feedback	Erlang Multi-Visit Path Traversal	Distributed mesh routing prevents single-node bottlenecking	5_n_node_feedback/n_node_with_feedback.py

---

Repository Structure

Attacks-Delay/
├── 1_one_node_destruction/        # Case 1: Single-node pre-service destruction
│   ├── one_node_packet_attack.py
│   ├── results_destruction.json
│   └── destroy_no_service_plot_reps200.png
├── 2_one_node_modification/       # Case 2: Single-node post-service modification
│   ├── one_node_packet_modification.py
│   ├── results_modification.json
│   └── plot_reps50_warmup500_sim5000.png
├── 3_tandem_chain/                # Case 3: Multi-hop tandem chain
│   ├── tandem.py
│   └── corrected_tandem_simulation_N3.png
├── 4_n_node_feedforward/          # Case 4: General N-node feedforward network
│   ├── n_node_feedforward.py
│   ├── results_feedforward.json
│   ├── section_3_3_2_stability_regions.png
│   └── section_3_3_2_tandem_delay_vs_N_varying_p.png
├── 5_n_node_feedback/             # Case 5: N-node symmetric feedback mesh
│   ├── n_node_with_feedback.py
│   ├── results_feedback.json
│   └── section_3_3_1_sojourn_vs_N_varying_p.png
├── arxiv_package/                 # Standalone arXiv submission bundle (main.tex + figures)
├── letter/                        # Cover letters and editorial correspondence
│   ├── cover_letter.tex
│   └── letter.tex
├── paper_ieee.tex                 # IEEE Transactions two-column manuscript
├── paper.tex                      # Standard single-column full research manuscript
├── ARXIV_METADATA.txt             # arXiv submission metadata (Title, Abstract, Categories)
├── SUBMISSION_GUIDE.md            # Step-by-step submission guide for arXiv and IEEE ToN
├── requirements.txt               # Python package dependencies
└── README.md                      # Project documentation

---

Installation & Setup

1. Prerequisites

• Python 3.8+
• TeX Live / MacTeX (optional, required only for compiling LaTeX paper manuscripts)

2. Environment Setup

# Clone the repository
git clone https://github.com/iliachry/Attacks-Delay.git
cd Attacks-Delay

# Create and activate a virtual environment
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

---

Reproducing Experiments & Generating Plots

Each case study can be executed independently. The scripts automatically compute theoretical analytical values, execute Monte Carlo SimPy discrete-event simulations across parameter sweeps, verify stability boundaries, and generate publication-quality figures:

Case 1: Single-Node Pre-Service Destruction

cd 1_one_node_destruction
python one_node_packet_attack.py

Outputs: results_destruction.json, destroy_no_service_plot_reps200.png

Case 2: Single-Node Post-Service Modification

cd 2_one_node_modification
python one_node_packet_modification.py

Outputs: results_modification.json, plot_reps50_warmup500_sim5000.png

Case 3: Tandem Queueing Chain

cd 3_tandem_chain
python tandem.py

Outputs: corrected_tandem_simulation_N3.png

Case 4: $N$-Node Feedforward Network

cd 4_n_node_feedforward
python n_node_feedforward.py

Outputs: results_feedforward.json, section_3_3_2_tandem_delay_vs_N_varying_p.png, section_3_3_2_stability_regions.png, section_3_3_2_tandem_throughput_vs_N.png

Case 5: $N$-Node Symmetric Feedback Mesh

cd 5_n_node_feedback
python n_node_with_feedback.py

Outputs: results_feedback.json, section_3_3_1_sojourn_vs_N_varying_p.png

---

Compiling Manuscripts & Publication Materials

Compiling LaTeX Papers

# Compile IEEE Transactions format (two-column)
pdflatex -interaction=nonstopmode paper_ieee.tex
pdflatex -interaction=nonstopmode paper_ieee.tex

# Compile Full Manuscript (single-column)
pdflatex -interaction=nonstopmode paper.tex
pdflatex -interaction=nonstopmode paper.tex

Packaging for arXiv Submission

The self-contained arXiv submission bundle can be generated directly:

cd arxiv_package
tar -czvf ../arxiv_submission.tar.gz main.tex *.png

See SUBMISSION_GUIDE.md and ARXIV_METADATA.txt for submission details.

---

Citation

If you find this codebase or theoretical framework helpful in your research, please cite:

@article{chrysovergis2026adversarial,
  title   = {Mathematical Modeling and Performance Analysis of Multi-Node Queueing Networks Under Adversarial Attacks and Timeout-Driven Retransmissions},
  author  = {Chrysovergis, Ilias},
  journal = {arXiv preprint},
  year    = {2026}
}

---

License

This project is licensed under the MIT License - see the LICENSE file for details.