Research On Probabilistic Routing Algorithm Based On Congestion Level In DTN

With the rapid development of wireless mobile communication,many new networks are emerging,such as interplanetary networks,Ad hoc networks,military networks,wireless sensor networks,which are different from traditional network based on the TCP/IP protocol,the path of end to end doesn't exist,and the network is intermittent connection,changeable network topology,long variable delay,asymmetric data rate and so on.In order to implement the special network communication,at the 2003 SIGCOMM International Conference,a new network architecture was proposed by DTN researchers such as Kevin Fall,it was called the Delay Tolerant Network(DTN),the new network layer was added between transport layer and application layer,and it was called bundle layer,through this layer can complete forwarding message to destination node by the mechanism of storage-carry-forward.There are two types of routing algorithms in the DTN,where one is the single copy routing algorithm and the other is the multi-copy routing algorithm.The multi-copy routing algorithm is implemented by increasing the number of replicas of message,which allows the same message to be forwarded to the relay node repeatedly,and finally submitted to the destination node through multiple paths.In the aspect of delivery rate,the multi-copy routing algorithm is bigger than the single copy routing,and the emphasis of research is the multi-copy routing algorithm.Therefore,in the restricted network environment,the effective multi-copy routing algorithm is of great significance.The main work of this thesis is as follows:Firstly,the probabilistic routing algorithm is based on historical information and probability of transfer.The node with high encounter probability to the destination node is chosen as the relay node,but it does not take historical connection time into account.If the historical delivery probability is high but each connection time is short,it may also lead to the message submitted unsuccessfully.Therefore,this thesis presents a new submission of expected value,which not only considers the encounter probability,but also increases the connection time of each encounter.Secondly,when the encounter probability of the relay node to the destination node is greater than the encounter probability of the current node to the destination node,the message is submitted to the relay node,but the congestion level of the relay node is not taken into account.When the relay node is congested,even though the encounter probability of the relay node to the destination node is large,the relay node receiving the message may lead to the occurrence of greater congestion or discard the received packets directly.Therefore,this thesis presents the concept of congestion level of node,which increases the judgment of congestion level of relay nodes.In accordance with the new delivery expectation and the judgment of the congestion level of the relay node,an improved probabilistic routing algorithm based on congestion level of node is proposed,which improves the network performance such as the delivery rate and average latency of the network.An improved probabilistic routing algorithm based on congestion level of node,this thesis adds the corresponding congestion detection mechanism and congestion control mechanism.When forwarding messages to the relay node,the message with the higher discarding policy factor D is given priority to forward,If the relay node is congested,the message with the smaller forwarding factor D are discarded in order to free up enough buffer space for the new messages.Comparing with the probabilistic routing algorithm based on congestion level of node without congestion control,which improves the delivery rate and reduces the overhead ratio.In this thesis,we use the ONE simulation software to simulate the proposed algorithm based on congestion level of node and congestion control strategy respectively,and make an analysis with other routing algorithms.