A Lightweight Trust Mechanism and Overhead Analysis for Clustered WSN

Abstract

Unrealistic trusted environment is the basis upon which sensor network's safety measures are built. The existing trust models for wireless sensor networks (WSNs) are not suitable for resource supplies and have high computation overhead. Our model, a lightweight trust mechanism and overhead analysis for clustered WSN (LWTM) has been designed on the dynamic nature of actual trust-building mechanism in order to meet the resource limitations of tiny sensor nodes. Here, a general framework of a realistic trust model for a large-scale clustered WSN is presented with a focus on protecting such system from various malicious attacks. In order to detect the malicious behaviour of nodes, malicious nodes up to 50% of total 500 nodes are intentionally incorporated in the proposed network. A trust metrics has been introduced here to highlight the important tasks of a sensor node. Metrics are divided into separate priority groups for assigning different importance levels. A dynamic trust updating algorithm based on the trust metrics priority of the parameters for rewarding or penalizing the trust values of the nodes is also proposed to assure sudden drop and gradual rise of trust. In addition to this, to avoid misjudgement of aggregated trust calculation, a self-adaptive weighted method is also defined for trust aggregation. On the model, to minimize the computational overhead through extensive simulation, we have shown that our trust model has less communication and storage overhead as compared to other two contemporary trust models such as GTMS and LDTS. The proposed trust model also provides better resilience against vulnerabilities. The feasibility of our trust model has been tested with MATLAB.