Research Of Efficient Data Distribution Mechanism In Opportunistic Networks

With the fast development of the network communication technology, the demands ofusers for network data communication efficiency become higher and higher. Nowadays, it hasbecome a research hotspot for data communication on how to utilize network resources moreefficiently and improve the performance of data communication. Opportunistic network doesnot require a complete communication link between a source node and a destination node, itachieves data communication by taking advantage of the meet opportunity brought bymovement of nodes in the network. To meet the demands of users, a data distributiontechnology which combines network coding with routing provides an effective solution to theissue. Network coding technology allows intermediate nodes in network to perform theencoding operation during data transmission,not just storing and forwarding the datas theyhava received. Network coding can significantly improve the communication performance ofnetwork, such as improving network throughput and the success rate of the network datacommunication and so on.Based on the typical characteristics of data transmission in opportunity network, thisarticle studies a data distribution mechanism in the aspects of improving the networkthroughput,reducing the end-to-end delay and improving the the success rate of the networkdata communication.The major innovative achievements in this thesis include the followingaspects:(1)This thesis first proposes an efficient data dissemination mechanismCRLNC(Clustering-Based Random Linear Network Coding) combined with clustering andrandom linear network coding.The core idea is to divide the data into several clusters firstly,and then to divide each cluster into the same number of data blocks, the source node sends thedata blocks in each cluster, the intermediate nodes encode the data blocks with random linearcoding algorithm and then forward the encoded data blocks, the destination node usesGauss-Jordan elimination method to restore the data progressively after receiving the encodeddata blocks. This algorithm has effectively solved the low success ratio problem of big data transmission, simultaneously it has obtained a more superior network performance inreducing the network delay and increasing the network throughput.(2) In this data distribution mechanism, the thesis also studies a cache problem ofnodes.Due to the limited storage space of nodes, when it receives some encoded packets andits buffer space is full, a nodes cache mechanism is required to determine the storage anddropout of the encoded data packets. Aiming at the redundancy problem of node cachespace,a node caching strategy is proposed based on cluster number and linear correlation. Thecaching mechanism of nodes in the network can effectively reduce the end to end delay anddata redundancy.