Skip to content
Open access

Empirical DDOS Detection for Constrained IOMT Edge Devices: A Hardware-in-the-Loop Cloud Inference Framework with Session-Aware Evaluation

Oct 2026 · Journal of Intelligent Decision Making and Information Science · 0 citations · 34 references

Abstract

Medical edge devices are increasingly integrated into hospital workflows, enabling continuous patient monitoring while simultaneously exposing fragile hardware to Distributed Denial of Service (DDoS) attacks. Existing intrusion detection frameworks rely predominantly on simulation datasets that cannot reproduce authentic hardware exhaustion signatures on physical constrained devices. To address this gap, we constructed a Hardware-in-the-Loop (HiL) testbed in which a physical ESP32 microcontroller—fitted with a heartbeat sensor and buzzer to emulate a bedside telemetry node—is subjected to four categories of flood traffic: HTTP GET, UDP, TCP SYN, and ICMP. Attack traffic was generated from a single host machine using Scapy with randomised spoofed source IP addresses (pool: 192.168.1.10–99) to replicate distributed botnet network-layer signatures. The complete dataset comprises 31,570 one-second telemetry samples collected over 68 days across 107 distinct acquisition sessions. Network and device metrics are aggregated by a dedicated monitoring host and forwarded to a Google Cloud Platform (GCP) n2-standard-4 virtual machine for classification. Eight classifiers are evaluated under a session-aware GroupKFold protocol (k=5), ensuring that all samples from a continuous acquisition session are assigned exclusively to one fold, preventing temporal autocorrelation leakage. Random Forest achieves the highest session-held-out accuracy of 98.44% (95% t-distribution CI: 97.22%–99.67%) and recall of 0.9812 (CI: 0.9527–1.000), outperforming all deep learning models on this 31,570-sample corpus. Cloud-side inference latency is 2.8 ± 0.2 ms; end-to-end detection latency (aggregation + transmission + inference) is approximately 1.1–1.3 seconds. This study constitutes an IoMT-inspired proof-of-concept; clinical-scale validation across multiple regulated devices and hospital-grade network topologies is identified as essential future work.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.