Social Routing Based On Location Preference Prediction In DTN

The traditional network is based on the TCP/IP protocol and needs to have a continuous end-to-end path.With the rapid development of mobile communication,such a network can no longer meet the demand of people to communicate anytime and anywhere.For example,in some special environments,such as deep space communications,war sites,disaster sites or remote villages these are challenging environments,there will be a series of problems such as link intermittent connection,network topology,long variable delay,and asymmetric data rate.As a result,many emerging networks continue to emerge,such as interstellar networks,land mobile Ad hoc networks,military networks,wireless sensor networks,and DTN(Delay Tolerant Network).Delay Tolerant Network architecture is a layer of bunching layer(Bundle layer)between the transport layer and the application layer,through the"Store-Carry-Forward" way to achieve the delivery of the message.The common routing methods in DTN are flooding-based routing,coded-based routing and forwarding-based routing.Each routing algorithm aims at improving delivery probability,reducing delay and overhead,and ultimately improving network performance.In DTN based on social network,mobile devices are usually carried and controlled by people.A mobile device can be considered as a mobile node.Therefore,social network-based DTN will exhibit a certain social relationship.In many routing algorithms,the nodes' positive social characteristics,such as centrality,friendship and community,are used to design the routing of DTN.The proposed Simbet or Bubble routing uses a small-world model that takes into account sociological and economic theories and extends DTN routing.In this thesis,we propose a social routing based on location preference in DTN,which combines the characteristics of social routing and delay tolerant networks,uses the basic principle of distributed waiting routing,dynamically distribute copies of messages.In this thesis,the main research is DTN architecture,several types of routing algorithms,efficient routing between different data 5 in current algorithm,packets are forwarded gradually through nodes with higher probability of visiting the destination node or area.However;these nodes are usually limited,the mobility of the destination node also makes the forwarding have a certain degree of blindness,through the analysis of this thesis,nodes move between different areas,so a certain number of copies are forwarded within each area,which can increase the probability of encounters between packets and destination nodes,according to the similarity of the two nodes,the scope of the message is increased in the network.In order to select better relay nodes,effectively enlarging the number of copies of a message in the network,this thesis proposes a routing algorithm for transferring between different landmarks by dividing the area,generally completed the following work:(1)The area of DTN is divided according to certain rules and divided into several smaller areas.Using the regularity and similarity of nodes,combining Markov model,distribute copies to the interior of each area based on a certain distribution strategy,proposes social routing based on Markov prediction.(2)According to the "strong and weak relationship" of information diffusion in social networks,the copies are dynamically distributed,so that the copies of the messages are spread as quickly as possible in the network,and the similarity of the encounter nodes is calculated using cosine similarity,propose location-based routing algorithm in DTN.(3)Use The ONE simulator to simulate the traditional routing and the proposed routing algorithm,and analyze the results.In summary,the social routing based on location preference prediction in DTN proposed in this thesis improves the probability of packet delivery,and reduces the average latency,which has certain theoretical significance and useful value.