Trust-based protocols for secure collaborative routing in wireless mobile networks
Usman, Aminu Bello
MetadataShow full metadata
In the context of Mobile Wireless Networks, researchers have proposed different collaborative routing schemes with fundamental objectives to maximise packet delivery, and minimise latency through improving peer routing decisions. One promising approach adopted by researchers to improve peers routing decisions is a trust-based routing mechanism. Trust-based routing protocols have several advantages including a better routing distribution strategy between wireless mobile peers and reliable forwarding decisions. The trust-based protocol enables peers in the network to dynamically forward data to a corresponding reliable and trustworthy peer for routing decisions and next peer selection strategy which can preserve the peers’ routing resources. This reduces the queuing delay and prevents routing attacks in the network. Along with several advantages of trust-based forwarding protocols, however, most of the current trust-based protocols of wireless mobile networks enable peers to make routing decisions based on trust relationships that manifest among peers but with less concern about the peers’ attributes. This further makes the design of efficient, collaborative routing protocols a challenging task due to the high dynamics of peers’ characteristics and mobility. In this study, we developed an efficient routing decision strategy to provide efficient, secure and higher quality communication between wireless mobile devices. Firstly, we explored the properties of complex networks and their impacts on trust and reputation propagation and evaluations, and presented the network performance analytical metrics for the design of an efficient trust-based forwarding protocol. Based on the theoretical, analytical and simulation studies observed in the study, we understand that the transitive contacts between the peers can be metrics elements of a trust-based routing protocol forwarding strategy. Secondly, we designed a Dynamic Attributes Trust Model for Efficient Collaborative Routing (DATM): a trust-based scheme to enforce collaborative behaviour in wireless mobile networks taking into consideration the peers’ attributes for an efficient routing scheme. DATM is a generic mechanism that can be integrated into any packet forwarding strategy or network management function to enable peers in the network to identify reliable, trustworthy peers for routing handling. Thirdly, based on the empirical evidence we exploited the existence of similar transitive contact patterns in the real data sets of Mobile Social Networks (MSNs). Therefore, we utilised the identified transitive connectivity properties to propose a new transitive data forwarding strategy which considers the similarity of transitive connectivity between peers for a data forwarding strategy. Our proposed transitive forwarding strategy gives preference to the peers with increasing transitive connectivity similarity to the messages’ destination to improve the chances of message carriers delivering the message to the correct destination. Finally, we proposed a model for evaluating the probability of the peers’ forwarding and receiving potentials for local trust evaluations based on peers’ forwarding and receiving potentials. In this regard, we introduced a correlation matrix for evaluating peers’ global trust values to improve the routing performance and mitigate the malicious effect of attacking peers in the network. Then we leveraged our proposed attributes similarity analytical model and transitive contact model for the design of an efficient forwarding strategy between peers in the network. The advantage of the proposed Forward-Watcher approach is that through routing history created by the peers, it can statistically predict the future behaviour of their corresponding routing partners. The concept can also serve as an underpinning and cornerstone for developing self-cooperative protocols in wireless mobile networks using statistical analysis and correlation matrices.