Temporal Distance Based Reachability Analysis And Its Application In Opportunistic Networks

With the rapid development of mobile Internet,communication costs and mobile network bandwidth pressure increase meanwhile.Opportunistic networks,as a special kind of mobile ad hoc network,can be deployed cheaply and flexibly and reduce the users' heavy burdens of mobile data fee.However,the routing of opportunistic networks is still a challenging problem and plenty of research in multi-duplicate routing and assistant information based routing are conducted.The temporal distance,expressing the time costs of the foremost journey between node pairs,is an important indicator of reachability of opportunistic networks.This thesis analyses the reachability of opportunistic networks on the basis of temporal distance and explores the characteristics of distribution and evolving of temporal distance.Two new routing mechanisms are designed from the perspectives of message delivery probability and temporal distance evolving.These routing mechanisms reduce latency of delivered messages and improve the ratio of delivered messages.The thesis also implements an application of content distribution system based on the new routing.The main work includes:(1)Analyzing the reachability of opportunistic networks from the perspective of temporal distance.This thesis proposes an algorithm to calculate the temporal distance based on the critical journey evolving graphs,and analyzes the reachability of the opportunistic networks from the perspective of the distribution characteristics of temporal distance,symmetry and time evolving characteristics.The results show that the temporal distance of most node pairs in the opportunistic network instances obeys the exponential distribution.In the case of the given appropriate tolerance error,the temporal distances of a node pair in opposite directions have a good symmetry.There is almost no white noise sequence in the temporal distance sequences,and most of their first order differential sequences are stationary.(2)Designing a routing mechanism based on message delivery probability.The regularity of temporal distance is applied to the design of opportunistic networks routing.Based on the distribution characteristics of temporal distance,the distribution parameters are estimated to calculate the delivery probability of messages with a given TTL(time to live)value.This routing mechanism uses a single copy to forward messages.For the purpose of improving the performance under the condition of insufficient symmetry,the parameter estimation of the temporal distance distribution is sent with the application message to the destination node.The simulation platform,ONE(Opportunistic Network Environment),verifies the latency and the ratio of successfully delivered messages under the routing mechanism.Comparing with other routing mechanisms such as PROPHET,a higher success rate of messages and a smaller average latency of delivered messages are obtained.(3)Designing a routing mechanism based on temporal distance prediction via time series.This routing predicts the temporal distance between the nodes at the current moment through time series method as the criterion of routing decision.Comparing to the routing based on message delivery probability,the time-dependent characteristics of temporal distance are considered,and there is no need to ascertain the distribution of temporal distance.The simulation results show that when TTL is not too short,this routing performs better in the latency and ratio of successfully delivered messages than the routing based on message delivery probability.(4)Applying routing based on temporal distance to content distribution system.The thesis implements a routing system on the Android platform decoupling the upper application and the underlying opportunistic network routing.Concise APIs are provided for upper application development and the APIs are used to implement a content distribution application during the periods of delivering content and forwarding interests.It significantly improves the flexibility and efficiency of path selection in content distribution.