The Research Of Spray And Wait Routing With Probabilistic Choose In DTNs

With the development of computer and communication technology in a high speed, many telecommunication equipment need to configure mobility ability, such as IPN, Wild animals tracking network, Battlefield military network, etc. The successful architecture and supporting protocols of today's Internet operate poorly when faced with operating environments characterized by very long delay paths and frequent network partitions. That network architecture have their specialized protocols, which is the Bundle protocol not the TCP/IP protocol. To achieve interoperability between them, the searchers had proposed a network architecture and application interface structured around optionally and reliable asynchronous message forwarding, with limited expectations of end to end connectivity and node resources. This architecture is called delay tolerant networks (DTN).This paper is mainly discuss the protocols, Many researchers have been start of the special research of the DTN network routing algorithms before, More representative is Ferry Routing, Epidemic Routing, Probabilistic routing and Spray and Wait routing. Among those routing arithmetic in DTN, the Ferry Routing is a single-copy routing, which has low delivery probability, high delay and litter network traffics. Epidemic Routing and Probabilistic routing can provide higher delivery probability and reducing delays, they cost too much network traffics. Spray and Wait routing can reduce the network traffic, but it has low delivery probability and high delay. In our paper, we combine the Spray and Wait routing and Probabilistic routing, we put the Probability choose in Spray and Wait routing. The simulation results we got indicate that this method can get higher delivery probability and lower delivery delay and lower network traffics. The paper's main methods and works as follows.(1) When spray packet in the original Spray and Wait routing, the choice of relay node has no rule, the choice of its relay node is the node that physical distance recently. If the relay nodes cannot be passed to the destination node immediately, the message will get the high delay and reduce the message delivery probability. In this paper, we put the Probability choose in Spray and Wait routing, every node has a probabilistic vector, when the two nodes meet, they compare their vector firstly, and when the vector is bigger than the other's, they will spray packet, therefore, the node has high probability will carry much packets.(2) The improvement of delivery message. When spray packet in the original Spray and Wait routing, the number of spray packet has two ways. The one method is source spray, which spray only one packet to relay node. The other method is Binary Spray and Wait, which spray half number packet to the relay node. In this paper, the number of spray packet is has something to do with the probabilistic vector, which is the product of probability and the number of packets in one node. Therefore, the node has high probability will get much packets.(3) The introduction of the Delivery Utility index. In the simulations, we introduce a new performance index, which is Delivery Utility, which can better assess the performance of the algorithm. We use the ONE simulator to build the human network. Then we simulate the routing algorithm, and compare the results with other routing algorithms.