Trust Management Framework For Fog Computing

Abstract

Fog computing is the new way forward to prevent network traffic outburst the exhausted IoT devices. It has moved the computation away from the central cloud to various logical points along its path, what used to be a direct connection of an IoT device to the central cloud now passes through a series of fog nodes, processing data along the way before it reaches the cloud. Few of the key highlights of deploying Fog include efficient use of bandwidth, increased reliability and reduced latency. The ease of deployment and flexibility of fog has attracted many, but various security and privacy issues hinder its vast deployment by major platforms. One such open issue if regarding the trust establishment, trust is the level of confidence that an object will behave in an acceptable manner. This thesis explores the requirements for establishing trust in fog environment, and through extensive literature review determines the similarities between SIoT and fog, as well as the trust requirement necessary for it. To mitigate the security and privacy issues of fog, we propose a two-way trust management scheme based on Bayes model, which allows both the service requestor and the service provider to validate the authenticity of each other before connecting. It is capable of effectively stopping a node from connecting with a malicious node. It is evaluated by simulating the system in Netlogo: an agent-based network simulator. The system shows resilience against trust-based network attacks and extensive evaluation shows that it has high accuracy and fastest convergence