A Clustering Routing Method Based on Semi-Markov Process and Path-Finding Strategy in DTN
E Wang · Chinese Journal of Computers · 2015
In delay tolerant network,we have proposed a clustering routing method based on semi-Markov process and path-finding strategy(CRSMP),which considers not only the nodes' social properties but also the time and maximum probability of node contacts.This method firstly clusters nodes according to path similarity,then predicts node contact probabilities at some point in the future by applying semi-Markov process and determines a collection of nodes used in the routing on the basis of cluster or clusters containing the source node and the destination node.Lastly,dynamic routing tables related to the current time are obtained by executing path-finding algorithm.Simulation results show that the delivery rate of CRSMP is much higher than DirectDeliveryRouter,FirstContactRouter and SimBetRouter which are also single-copy routing schemes under the condition that the size of caches is small,and the delivery rate of CRSMP is also higher than Spray and Wait,Epidemic and Prophet which are multi-copy routing schemes,andCRSMP with 10M caches is similar to Epidemic with 500M caches in routing performance.Furthermore,experiments with the real data set are done,and CRSMP still keeps better routing performance.