Research On Routing Strategies Based On Delivery Probability And Message Redundancy In Delay Tolerant Networks

Delay Tolerant Network (DTN) is a general, message oriented and reliable networkarchitecture. It is used to connect heterogeneous field networks which are characterizedby high delay, low rate of data transfer, intermittent connection and high ratio of errorcode. It is a novel mobile ad-hoc network which communicates by the opportunity whichis brought in by nodes’ movement in the network, and design of the network architectureis in the aim of addressing the data communication problem in frequently intermittentnetwork. In this kind of network, on the reason of nodes’ frequent movement, sparsedistribution of nodes and limitation of nodes’ energy and storage, the whole network isdivided into several parts. Thus, the source node and destination node are distributed indifferent fields, so traditional wireless ad-hoc network protocol cannot work in delaytolerant network. Hence, the architecture of delay tolerant network is different frominternet. Recently, the research topic of delay tolerant network is focused on networkarchitecture, routing strategy, congestion control strategy and security issues etc.Based on PRoPHET routing protocol, we have proposed a novel routing algorithmwhich based on hybrid of message delivery probability and redundancy (HPR).According to algorithm PRoPHET, the history information of nodes’ encounter decidesnode’s delivery probability of message. Every node in network will maintain a deliveryprobability table, when two nodes meet each other, they will compare each other’sdelivery probability table to decide whether to forward message to the met node.However, in this algorithm, it doesn’t consider the impact of contact duration betweennodes when two nodes meet each other, so this algorithm can only achieve a wellperformance on the provision of once two nodes meet each other the message can betransferred successfully. While, in real network environment this provision will fall indilemma. Thus, on the consideration of impact of contact duration on message delivery,we consider contact duration and encounter frequency together to calculate deliveryprobability. In addition, in order to control network overhead, the maximum number ofmessage copies is designated in network and the task of copy generation is distributed in binary mode.In order to evaluate performance of the routing algorithm, we have implemented thisalgorithm in simulator ONE (Opportunistic Network Environment). By the experimentalresults, the algorithm will increase message’s delivery efficiency and reduce networkoverhead significantly.