Senior Project 2025–2026

Intelligent
Identity & Access Management
with Active Defense

A five-node closed-loop distributed cybersecurity system that integrates honeypot-driven attack intelligence with machine learning-based risk authentication to protect against identity-based threats in real time.

98.87%
Classification Accuracy
4,710+
Attack Events Captured
A+
SSL Labs Rating
🧠 IAM
🖥️ Frontend
📊 Risk API
🔬 ML Pipeline
🪤 Honeypot
🛡️ WAF
🚀 Next.js ⚡ FastAPI 🔐 Methaq IAM 🪤 Cowrie Honeypot 🤖 Isolation Forest 🌲 Random Forest 🛡️ ModSecurity 🔒 TLS 1.3 🐳 Docker 🐘 PostgreSQL 🚀 Next.js ⚡ FastAPI 🔐 Methaq IAM 🪤 Cowrie Honeypot 🤖 Isolation Forest 🌲 Random Forest 🛡️ ModSecurity 🔒 TLS 1.3 🐳 Docker 🐘 PostgreSQL
🏗️ Architecture

Five-Node Distributed System

Each node serves a specific role in the closed-loop intelligence pipeline, working together to detect, classify, and respond to identity-based threats.

🖥️
Node 1

Methaq IAM Server

Identity provider with custom RiskScoreAuthenticator SPI enforcing risk-based authentication decisions (allow/challenge/block).

Keycloak Java SPI OIDC PKCE
🌐
Node 2

Frontend & WAF

Next.js demo banking application protected by Nginx reverse proxy with ModSecurity WAF and OWASP Core Rule Set.

Next.js Nginx ModSecurity OWASP CRS
📊
Node 3

Risk API & ML Engine

FastAPI-based risk scoring service with ensemble ML model combining Isolation Forest and Random Forest classifiers.

FastAPI Python Scikit-learn SMOTE
🪤
Node 4

Honeypot

Cowrie SSH/Telnet honeypot capturing live attack data on owned infrastructure with JSONL event logging.

Cowrie SSH Telnet JSONL
🔄
Node 5

ML Retraining Pipeline

Continuous retraining pipeline that refreshes the ML model within 18 minutes of new attack data capture.

Python ONNX SMOTE Cron
✨ Features

Why Methaq?

Methaq addresses the fundamental gap between attack detection and identity enforcement with a closed-loop intelligence architecture.

🔄

Closed-Loop Intelligence

Attack data captured by the honeypot flows to the ML engine for classification, and risk scores drive authentication decisions — creating a self-improving security architecture.

🤖

ML-Powered Risk Scoring

Ensemble model combining Isolation Forest (0.4 weight) and Random Forest (0.6 weight) with SMOTE oversampling for 98.87% classification accuracy.

🎯

Risk-Based Authentication

Three-tier authentication decisions: Allow (score < 0.5), Challenge/MFA (0.5 ≤ score < 0.8), Block (score ≥ 0.8) based on real-time threat intelligence.

🛡️

Multi-Layer Defense

WAF with OWASP CRS, TLS 1.3 encryption, Zero-Trust architecture, fail2ban intrusion prevention, and LUKS disk encryption across all nodes.

📈

Real-Time Adaptation

The system continuously retrains its ML model within 18 minutes of new attack data capture, ensuring near-real-time adaptation to emerging threats.

🔍

Comprehensive Monitoring

Real-time audit logs, security events dashboard, session revocation, and device/location anomaly detection for complete visibility.

📊 Results

Proven Performance

Evaluation demonstrates the effectiveness of the closed-loop architecture across multiple metrics.

98.87%
Classification Accuracy
4,710+
Attack Events Captured
< 18min
Retraining Latency
118ms
Risk API Response
A+
SSL Labs Rating
5/5
Pen Tests Blocked
0
OWASP ZAP Findings
📚 Resources

Explore Methaq

Access all project resources, documentation, and downloads.

👥 Team

The Team Behind Methaq

A dedicated team of computer science students from the University of Hafr Al-Batin.

AA

Abdulrahman Al-Anazi

Project Leader & System Architect

Designed the five-node architecture, developed the RiskScoreAuthenticator SPI, and led all architectural review sessions.

MA

Mansour Al-Anazi

Application & OIDC Engineer

Developed the demo banking application with OIDC client integration and PKCE authorization code flow.

AA

Abdulmohsen Al-Anazi

Security & Frontend Engineer

Responsible for security configuration, WAF rule customization, SSL/TLS hardening, and vulnerability assessment.

AH

Abdullah Al-Harbi

Frontend & WAF Lead

Led the deployment of Nginx reverse proxy and ModSecurity WAF with OWASP CRS on Node 2.

HS

Hamed Salem Al-Anazi

Infrastructure & IAM Lead

Provisioned Hetzner Cloud infrastructure, deployed Methaq IAM server, and configured Caddy reverse proxy with TLS 1.3.

FH

Faisal Al-Harbi

Machine Learning Engineer

Designed and trained the ML ensemble with SMOTE oversampling, achieving 98.87% classification accuracy.

AA

Abdullah Al-Anazi

Application Tester & Data Engineer

Designed comprehensive test plans, executed all testing phases, and built the data collection pipeline.

YA

Yousef Al-Anazi

Active Defense Engineer

Deployed and configured the Cowrie honeypot, managed collection of 4,710+ attack events, and designed the threat intelligence feed.

Ready to Explore?

Experience the Methaq system firsthand with our live demo, or dive into the technical documentation.