ReTREM: A responsibility based trust revision model for determining trustworthiness of fog nodes

Fog computing complements the Internet of Things (IoT)-Cloud paradigm by enabling local computation near IoT devices. Fog nodes support IoT applications such as health care monitoring, drone surveillance, and accident reporting. A minor breach or inaccurate results (caused when applications are placed at malicious fog nodes) can cause chaos or severe damage to users. Hence, it is of utmost vital that these applications must be processed at trustful fog nodes. In this work, we have proposed a responsibility-based trust revision model, named ReTREM, for determining and monitoring the trustworthiness of fog nodes. ReTREM categorizes fog nodes based on their response when they violate the promised service quality, i.e., their intention behind malicious behavior. A separate penalty mechanism is proposed for each category of fog node to discourage their malicious behavior. ReTREM considers two types of trust scores: TS1 and TS2. TS1 is calculated based on various metrics related to trust, such as reliability decay, data integrity, social rating, and feedback rating, while the fog broker gives TS2. Trust score (TS), i.e., the sum of TS1 and TS2, is revised based on various proposed penalty mechanisms. The proposed model rewards the fog node that does not violate the promised service. The proposed model is simulated and compared with the two variants of application placement, and it outperforms. The proposed model encourages fog nodes to provide the best service and increases the users' faith. The proposed model also tackles various trust-based attacks resulting in fair competition and opportunity among fog nodes.