Web-Based Botnet for Blocking Control Flow in Open-Source Medical Syringe Pump.

Area of Science:

• Biomedical Engineering
• Computer Science
• Cybersecurity

Background:

• Open-source medical systems and advancements in 3D printing and microcomputers (Arduino, Raspberry Pi) have transformed healthcare.
• This integration has introduced significant cybersecurity vulnerabilities within hospital networks.
• Internet of Things (IoT) medical devices are increasingly susceptible to network-based attacks.

Purpose of the Study:

• To present a proof-of-concept web-based botnet targeting a syringe pump within an IoT medical network testbed.
• To demonstrate the potential for botnets to disrupt the control flow of critical medical devices.
• To provide a publicly available dataset for cybersecurity research on open-source medical systems.

Main Methods:

• Development of a lightweight, resource-efficient, and rapidly deployable web-based botnet.
• Creation of a dedicated IoT medical network testbed for botnet attack simulation.
• Development and application of a novel feature selection methodology for botnet attack detection.
• Comparative analysis of various machine learning algorithms for attack identification using network traffic data.

Main Results:

• The developed botnet successfully demonstrated potential disruptions in the control flow of a syringe pump.
• The feature selection technique achieved over 99% accuracy in detecting botnet attacks on the testing dataset.
• The methodology successfully identified 63,380 out of 63,382 malicious attack instances.

Conclusions:

• Open-source medical systems, while innovative, present critical cybersecurity challenges.
• The developed botnet and dataset offer valuable resources for cybersecurity research in healthcare IoT.
• The proposed feature selection methodology provides an effective strategy for detecting botnet attacks in medical networks with high accuracy.