DEVELOPMENT OF A PRIVACY AIDED TRUST ROUTING ALGORITHM BASED ON SOCIAL SIMILARITY IN OPPORTUNISTIC NETWORK

DEVELOPMENT OF A PRIVACY AIDED TRUST ROUTING ALGORITHM BASED ON SOCIAL SIMILARITY IN OPPORTUNISTIC NETWORK

Abstract:

This research work presents the development of a privacy aided trust routing model using social similarity in an Opportunistic Network (OppNet). OppNet is a delay tolerant network where link is highly unpredictable. It is an ideal solution in situations where the deployment of wired and conventional wireless networks is difficult or practically impossible. However, OppNet is faced with several challenges. Most of these challenges are related to routing, buffer management or security. The most promising amongst the several approaches by researchers at addressing these challenges are the ones that explored social behavior of human beings. The mobile nature of OppNets however brought to the fore the need for a more efficient and effective security model. This work, therefore, intends to further strengthen privacy in an OppNet. This is done by developing a community-based privacy preservation scheme using a symmetric cryptographic model, based on a differential equation of a curvi-circular function by applying Galois theory. The key distribution is made possible by partial application of onion routing scheme. The developed privacy scheme is incorporated into trust routing based on social similarity (TRSS) to become Privacy Aided TRSS (PATRSS). This approach improves the security in terms of preservation of privacy of messages and that of the relay nodes without compromising other network performance indicators. Using metrics, such as delivery ratio, delivery cost and average trust performance, PATRSS is simulated on Opportunistic Network Environment (ONE) simulator. PATRSS out performed TRSS by 7.8%, 30.7% and 9.4% in terms of delivery ratio, delivery cost and average trust value respectively. Finally, the TRSS and PATRSS which are originally routed on spray and wait routing algorithm are implemented on Epidemic, PRoPHET and MaxProp. Using numbers of messages duplicated, those aborted and those delivered as metrics, the results are presented to clearly show the network performance, the improvements, and the costs. These demonstrate the effectiveness and the efficiency of PATRSS.

DEVELOPMENT OF A PRIVACY AIDED TRUST ROUTING ALGORITHM BASED ON SOCIAL SIMILARITY IN OPPORTUNISTIC NETWORK