Opportunity Network Routing Mechanism Study Based On Community Detecting

Intermittently connected mobile networks or opportunity networks are wirelessnetworks where most of the time there does not exist a complete path from the source tothe destination. There are many real networks that follow this model, for example,wildlife tracking sensor networks, military networks, vehicular ad hoc networks, etc. Inthis context, conventional wireless routing schemes fail, because they try to establishcomplete end-to-end paths, before any data is sent. In opportunity network, packets areforwarded to relay nodes with assistance of mobility, so-called "store-carry-forward"manner. Current routing scheme try there best to delivery packet but leading todefinitely network delay and are based on random movement model. However, aftersmall world social network model is proposed recently, conventional routing schemesare not effective in the new model.According to the features of social network, this article introduce the expectedmeeting time of nodes in the community movement model after researching theregulation of random direction mobility model. Then, this article analysis the movementepoch features of community movement model, estimate the expected meeting time slotduring each movement epoch, and forcast expected meeting delay between carrier anddestination node. Whereafter a expected meeting delay as a direction of routing schemeas well as congestion control strategy in proper opportunity is proposed.Finally thescheme is proved to have better performance after experiment, and estimating algorithmis confidently correct, and the network performance is get better.Because current community detecting algorithms in opportunity network havecommon in these: community edge is not clear, premises are strict, network topographyinformation is high redundant. So SACD(short for Similarity Aware CommunityDetecting) is proposed. This algorithm is based on similarity among nodes and labelpropagation scheme. According to the statistics of roaming epochs in communitymovement, the centrality of nodes can be evaluated accurately; moreover, the networkupdate correlative situation after controlling the label propagation with certainconstraint condition. And the whole network is divided into several communitieslogically while central nodes are also probed. Simulation results show that the method iseffective. The network is show to us with central nodes and neighbors clique. Finally CCR(short for Community Centrality Routing) is proposed based on the detectingalgorithm. This routing makes use of central nodes is more active among network, andassist other ordinary nodes with forwarding packets, nodes help each other amongneighbors in the same logical community. Simulation value analysis shows that CCRhas good network performance in social network, especially in the environment oflimited network resources, long TTL of packets, wide events interval of new packets.