Framework for Early Cyber Attack Detection Using ML Models Deployed On Fog Devices
K. Aditya, Prateek N Kamath, Yuktha Poral, Baddela Divya Mallika, Vadiraja Acharya · 2024
The growing Internet of Things (IoT) is set to link facets of life including intelligent structures, residences, and whole urban areas. A key factor propelling this growth originates from progress, in the semiconductor sector. IoT promises innovative services demanding extensive data centers and real-time processing capabilities. However, an advancement in technology also promises an advancement in the attack surface and the threats these devices will have to face. IoT security is an up-and-coming field, raising awareness of the need for security mechanisms for these constrained devices. Existing solutions are computationally heavy for a traditional constrained device in the Fog. With the limitless implementations of these devices, there is a need for a security solution that can be scaled to a large majority of IoT applications. This paper presents a Python-based Framework with a variety of security methods and functions aimed at the early detection of Cyber Attacks using multiple ML models in real-time. The Framework is tested on a Raspberry Pi 4 device deployed within the fog computation layer, functioning as an intermediary between edge and cloud layers. This fog device is subsequently linked to a network of IoT devices. A distinctive aspect of this Framework is its dual perspective, catering to both developers and general users.