TEAM is developing a framework that will enable the deployment of trusted edge-based multi-tenanted IoT network where applications are of varying criticality.
IoT devices will be widely deployed in the future. These IoT devices can be low on resources, perform many activities, and be running many applications of varied levels of risk. High risk applications will have more resources available at the edge of the network. The functionality of the IoT network is dependent on the reliability at the edge node. When an edge node is compromised, the properties of the applications may be violated. Understanding which edge node to trust for timely and correct execution becomes important. However, the trust information held by the IoT devices needs to be contextualised and may also be compromised.
Based on the threat model, TEAM is developing a contextualised trust model that is resilient to the identified threats. TEAM develops trust-based algorithms that assign jobs to edge nodes to ensure jobs are completed according to their requirements. The allocation information and execution profile is saved so that machine learning algorithms such as reinforcement learning can be executed offline on the data to predict successful allocations.