A two-way trust management system for fog computing

Fog computing is the next frontier of cloud computing since it can compute and store a massive amount of data generated by IoT devices near their sources. Indeed, transmitting all these data to the cloud will take up a huge amount of bandwidth. However, its features and flexibility of deployment make fog computing vulnerable to security and privacy attacks. The high-mobility support, dynamic environment, geographical distribution, location awareness, proximity to end users, and lack of redundancy are among these features that make the existing schemes not adequate to fog computing. Therefore, since trust management ensures security and privacy, we propose a two-way subjective logic-based trust management system that enables a service requester to verify whether a service provider can give reliable and secure services and lets the service provider check the trustworthiness of the service requester. Extensive evaluation of the system shows that the trust value of a node is accurate and converges in a very few trust computation cycles. The solution is also resilient to a large population of misbehaving nodes and it is able to thwart trust-based attacks successfully. Moreover, comparative analysis is made with different modification of the system and two baseline schemes, PeerTrust and EigenTrust. The comparison depicts that the two-way trust management proposals have lower overhead, allow a balanced load distribution, are effective in selecting the right service providers and are more accurate than the conventional one-way trust management systems. •Fog Computing is Vulnerable to security and privacy attacks.•Two-way trust management system ensures security and privacy.•Adaptive weighting increases trust accuracy of good fog nodes.•Behavior change of fog nodes is captured by the trust management system.•Two-way trust management outperforms one-way trust management system.