No abstract

No abstract

No abstract

Purpose -Mobile ad hoc networks rely on cooperation to perform essential network mechanisms such as routing. Therefore, network performance depends to a great extent on giving participating nodes an incentive for cooperation. The level of trust among nodes is the most frequently used parameter for promoting cooperation in distributed systems. There are different models for representing trust, each of which is suited to a particular context and leads to different procedures for computing and propagating trust. The goal of this study is to analyze the most representative approaches for mobile ad hoc networks. It aims to obtain a qualitative comparison of the modeling approaches, according to the three basic components of a trust model: information gathering, information scoring and ranking, and action execution. Design/methodology/approach -The paper identifies the different tasks required by a trust system and compares the way they are implemented when the system model itself is based on information theory, social networks, cluster concept, graph theory and game theory. It also provides a common nomenclature for the models. The study concentrates exclusively on the trust models themselves, without taking into account other aspects of the original articles that are beyond the scope of this analysis. Findings -The study identifies the main components that a trust model must provide, and compares the way they are implemented. It finds that the lack of unity in the different proposed approaches makes it difficult to conduct an objective comparison. Finally, it also notices that most of the models do not properly manage node reintegration. Originality/value -The best of our knowledge, the study is the first that uses information scoring and ranking as classification key. According to this key, approaches can be classified as based on information theory, clusters and social network theory, and cooperative and non-cooperative game theory. It also provides a common nomenclature for all of them. Finally, the main contribution of the paper is to provide an analysis of the most representative approaches and present a novel qualitative comparison.

No abstract