DESIGN AND IMPLEMENTATION OF A REALTIME CYBER-AWARE MONITORING SYSTEM FOR SMART FARM NETWORKS USING STRIDE THREAT MODELLING

Abstract

Agricultural systems are increasingly digitalised through the Internet of Things (IoT), enhancing productivity and precision while simultaneously heightening exposure to cyber threats. This paper presents the design and implementation of a real-time cyber-aware monitoring system for smart farm networks using the STRIDE threat-modelling framework. Unlike previous conceptual applications of STRIDE, each threat category— Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, and Elevation of Privilege—was operationalised into executable detection rules within a hybrid anomaly detection engine. The system combines deterministic logic derived from STRIDE with an Isolation Forest model (n_estimators = 200, contamination = 0.05, max_samples = 256) to enable adaptive zero-day threat detection and automated quarantine of compromised MQTT nodes. Using the Mendeley Smart-Farm dataset comprising approximately 3,000 sensor readings of temperature, humidity, and soil moisture, real-time streaming experiments were conducted under a 70:15:15 split. The system achieved Accuracy = 0.96, Precision = 1.00, Recall = 0.91, F1- score = 0.95, with a mean Detection Latency of 0.55 seconds and mean Time-to-Quarantine of 0.87 seconds. The findings confirm that integrating STRIDE with real-time detection and containment enhances resilience, auditability, and adaptive security within precision agriculture IoT environments.