🔐 DKM - Dynamic Key Management System

A sophisticated file transfer system with intelligent key rotation based on image similarity analysis and post-quantum cryptography. DKM automatically rotates encryption keys when consecutive images show significant differences, ensuring enhanced security for sensitive data transmission.

---

🎯 What is DKM?

DKM (Dynamic Key Management) is a secure client-server file transfer system that revolutionizes encryption key management. Instead of using static keys or time-based rotation, DKM analyzes the content similarity of consecutive images using Mean Squared Error (MSE) metrics and automatically rotates encryption keys when dissimilarity is detected.

💡 Why DKM?

Traditional encryption systems use:

• Static keys - vulnerable to long-term exposure
• Time-based rotation - arbitrary and potentially inefficient
• Manual rotation - error-prone and inconsistent

DKM uses content-aware rotation that:

• ✅ Adapts to actual data patterns
• ✅ Triggers rotation only when needed (similarity < 92%)
• ✅ Uses quantum-resistant ML-KEM-1024 for new key generation
• ✅ Ensures perfect forward secrecy with AES-256 encryption

---

✨ Features

🔑 Core Capabilities

• Intelligent Key Rotation - Automatic encryption key updates based on image similarity analysis
• Post-Quantum Cryptography - ML-KEM-1024 (NIST standardized) for quantum-safe key generation
• AES-256 Encryption - Military-grade encryption for all file transfers
• Image Similarity Detection - MSE-based comparison with configurable threshold (default: 0.92)
• DWT Processing - Discrete Wavelet Transform for image decomposition and analysis
• Robust Error Handling - Retry mechanisms with acknowledgment-based confirmation
• Real-time Logging - Comprehensive activity tracking for security auditing

🔒 Security Features

• SHA-256 checksums for data integrity
• Socket-level encryption with key derivation
• Automatic key synchronization between client and server
• Secure key encapsulation using ML-KEM

---

🛠️ Tech Stack

Component	Technology
Language	Python 3.8+
Encryption	AES-256 (CBC mode)
Post-Quantum Crypto	ML-KEM-1024 (Kyber)
Image Processing	scikit-image, OpenCV, Pillow
Wavelet Transform	PyWavelets (Haar wavelet)
Networking	TCP Sockets
Serialization	Pickle

---

🚀 Quick Start

Prerequisites

• Python 3.8 or higher
• Network connectivity between client and server
• Images in JPG, PNG, BMP, TIFF, or GIF format

Installation

1. Clone the repository

git clone https://github.com/AkshayKappala/dkm.git
cd dkm

2. Install dependencies

pip install -r requirements.txt

3. Configure network settings

Edit the server address in both client/client.py and server/server.py:

SERVER_ADDRESS = ('YOUR_SERVER_IP', 12345)

Running the System

Terminal 1 - Start the Server:

python -m server.server

Terminal 2 - Run the Client:

python -m client.client

Place your images in the sent/ directory before running the client.

---

📁 Project Structure

dkm/
├── client/
│   ├── client.py                    # Main client with key rotation
│   ├── client1.py                   # Basic client (no rotation)
│   ├── encryption/
│   │   ├── aes_encryption.py        # AES-256 encryption logic
│   │   └── dwt_processor.py         # Discrete Wavelet Transform
│   └── utils/
│       └── file_utils.py            # File I/O operations
├── server/
│   ├── server. py                    # Main server with key rotation
│   ├── server1.py                   # Basic server (no rotation)
│   └── decryption/
│       ├── aes_decryption.py        # AES-256 decryption logic
│       └── dwt_reconstructor.py     # Image reconstruction
├── shared/
│   ├── crypto_utils. py              # Shared cryptographic utilities
│   └── key_rotation_manager.py      # Image similarity & key rotation
├── ID_MSE. py                        # Image similarity computation
├── requirements.txt                 # Python dependencies
└── README.md                        # This file

---

🔧 How It Works

1. Image Similarity Analysis

def compare_images(image1, image2):
    # Convert images to float
    image1 = img_as_float(image1)
    image2 = img_as_float(image2)
    
    # Resize to same dimensions
    image2 = resize(image2, image1.shape)
    
    # Compute MSE and similarity
    mse = np.mean((image1 - image2) ** 2)
    similarity = 1 - mse
    return similarity

2. Dynamic Key Rotation Logic

The system compares consecutive images:

• Similarity ≥ 0.92: Keep current key
• Similarity < 0.92: Generate new key using ML-KEM-1024

if similarity_score < self.similarity_threshold:
    ek, dk = ML_KEM_1024.keygen()  # Generate quantum-safe keypair
    shared_key, ciphertext = ML_KEM_1024.encaps(ek)
    new_password = shared_key.hex()  # New encryption key
    return True, similarity_score, "Low similarity detected", new_password

3. Secure File Transfer Protocol

Client → Server: [Flag:  0x01] [Key Length] [New Key]
Server → Client: ACK

Client → Server: [Flag: 0x02] [Filename Length] [Filename] [Data Length] [Encrypted Data]
Server → Client: ACK

Client → Server: [Flag: 0x03] (End of Transfer)
Server → Client: ACK

---

⚙️ Customization Guide

Adjust Similarity Threshold

In shared/key_rotation_manager.py:

key_rotation_manager = KeyRotationManager(similarity_threshold=0.85)  # More sensitive

Change Encryption Password

In both client/client.py and server/server.py:

password = "your_secure_password_here"

Modify Server Port

SERVER_ADDRESS = ('192.168.1.100', 8080)  # Custom IP and port

Configure Socket Buffer Size

In client/client1.py:

client_socket.setsockopt(socket.SOL_SOCKET, socket.SO_SNDBUF, 131072)  # 128KB
client_socket.setsockopt(socket.SOL_SOCKET, socket.SO_RCVBUF, 131072)  # 128KB

---

🚢 Deployment

Production Considerations

1. Use Environment Variables for Sensitive Data

import os
password = os.getenv('DKM_ENCRYPTION_KEY', 'default_secure_password')
SERVER_ADDRESS = (os.getenv('DKM_SERVER_IP', '0.0.0.0'), 12345)

2. Enable TLS/SSL for Socket Communication

import ssl
context = ssl.create_default_context(ssl.Purpose.CLIENT_AUTH)
context.load_cert_chain(certfile="server.crt", keyfile="server.key")
secure_socket = context.wrap_socket(server_socket, server_side=True)

3. Run as a Service (Linux)

# Create systemd service file
sudo nano /etc/systemd/system/dkm-server.service

[Unit]
Description=DKM Server
After=network.target

[Service]
Type=simple
User=dkm
WorkingDirectory=/opt/dkm
ExecStart=/usr/bin/python3 -m server.server
Restart=always

[Install]
WantedBy=multi-user.target

sudo systemctl enable dkm-server
sudo systemctl start dkm-server

4. Docker Deployment

FROM python:3.9-slim

WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY . . 
EXPOSE 12345

CMD ["python", "-m", "server.server"]

docker build -t dkm-server .
docker run -d -p 12345:12345 --name dkm-server dkm-server

---

🐛 Troubleshooting

Connection Refused

# Check if server is running
netstat -tulpn | grep 12345

# Verify firewall rules
sudo ufw allow 12345/tcp

Decryption Errors

• Ensure client and server use the same initial password
• Check logs for key rotation acknowledgments
• Verify network stability during key transmission

File Not Received

• Confirm sent/ directory contains valid image files
• Check server logs for specific error messages
• Increase socket timeout: client_socket.settimeout(10)

Import Errors

# Reinstall dependencies
pip install --upgrade -r requirements.txt

# Install missing kyber-py
pip install kyber-py

Image Similarity Not Triggering Rotation

• Lower the threshold: similarity_threshold=0.80
• Ensure images are significantly different
• Check that files are valid images (not corrupted)

---

📊 Performance Metrics

Operation	Average Time
Key Generation (ML-KEM-1024)	~10ms
AES-256 Encryption (1MB file)	~50ms
Image Similarity Calculation	~100ms
File Transfer (1MB over LAN)	~200ms

---

🔬 Example Usage

Basic File Transfer

# Place images in sent/ directory
sent/
├── image1.jpg
├── image2.jpg
└── image3.png

# Run client
python -m client.client

Output:

2025-12-16 10:30:00 - INFO - Connecting to server at 192.168.233.129:12345
2025-12-16 10:30:01 - INFO - Processing file:  image1.jpg (Attempt 1)
2025-12-16 10:30:01 - INFO - Key rotation decision: False (Reason: First image received)
2025-12-16 10:30:01 - INFO - File image1.jpg encrypted successfully. 
2025-12-16 10:30:02 - INFO - Processing file: image2.jpg (Attempt 1)
2025-12-16 10:30:02 - INFO - Key rotation decision: True (Reason: Low similarity 0.7843 < 0.92)
2025-12-16 10:30:02 - INFO - Rotating key for file: image2.jpg

---

🤝 Contributing

Contributions are welcome! Areas for improvement:

• Support for non-image file types
• Web-based dashboard for monitoring
• Multi-client support with threading
• Additional similarity metrics (SSIM, perceptual hashing)

---

📞 Contact

Developer: Akshay Kappala
GitHub: @AkshayKappala
Repository: github.com/AkshayKappala/dkm

---

📜 License

This project is open source. Please add a LICENSE file to specify terms.

---

🙏 Acknowledgments

• NIST - For ML-KEM-1024 post-quantum cryptography standard
• scikit-image - Image processing capabilities
• PyCryptodome - Robust AES implementation

---

Built with 🔐 by Akshay Kappala

Securing the future with quantum-resistant cryptography