Cache Management Routing Algorithm Based On Node Estimated Utility And Message Delivery Rate In DTN

Delay Tolerant Network(DTN)is an end-to-end network architecture that uses a "carry-store-forward" mechanism for message transmission.Nowadays,DTN is used in many fields,such as in-vehicle network,satellite communication and disaster emergency response.In DTN,the mobility model is one of the important factors that affect the routing performance.However,most researchers have only focused on improving the routing algorithm while ignoring the impact of the mobility model on the routing algorithm.Studies have shown that routing algorithms exhibit different routing performance under different mobility models.Furthermore,the traditional Spray and Wait algorithm always selects nodes blindly,ignoring the nodes’ own merits and demerits,resulting in unreasonable copy allocation.Based on a Map-Based Movement model,this paper proposes a Node Estimated Utility-based and Forwarding Probability Spray and Wait routing algorithm(FPEU-SAW).In the Spray phase,the algorithm uses exponential smoothing to estimate the effective change rate of node neighbors and the effective change rate of node messages under the Map-Based Movement model.Based on this,the algorithm calculates the node estimated utility value and adaptively allocates message replicas.In the Wait phase,the FPEU-SAW algorithm considers the properties of the internode interval time between nodes following the exponential distribution and the connection time between nodes following the power-law distribution under the MBM model.It calculates the probability of node encounters and the probability of successful message delivery.Then,it introduces the concept of message delivery rate and selects relay nodes based on the message delivery rate.In addition,proposed a cache management strategy that utilizes the node message quality,defining the message quality by combining the number of message forwarding times,the message transmission time,and the remaining message lifetime.The priority of message deletion is determined based on the message quality.The message forwarding policy related to the buffer state between nodes needs to be adopted to forward messages during the Wait phase.The main work is as follows:(1)The article proposes a Node Estimated Utility-based and Forwarding Probability Spray and Wait routing algorithm(FPEU-SAW).The effective neighbor change rate is defined,and the exponential smoothing method is used to predict a node’s ability to increase the number of neighbors in the next moment.Choosing nodes with higher effective neighbor change rates is more likely to forward messages to other nodes,thus increasing the spread of messages in the network.At the same time,the paper defines the effective message change rate for message prediction to estimate the amount of messages that a node will forward in the next moment.Nodes with higher effective message change rates have stronger capabilities to forward messages,and selecting such nodes as relay nodes can effectively alleviate congestion and improve message delivery rates.Combining the neighbor prediction effective change rate and message prediction effective change rate,this paper defines the node prediction utility value.The copy allocation based on the utility value can increase the degree of message diffusion in the network,thus increasing the delivery rate of messages.Secondly,considering that the connection time between nodes follows a power law distribution and the contact time between nodes follows an exponential distribution under the MBM model.In this paper,we defines the message delivery rate to improve the Wait phase.Finally,a comparative experiment was conducted using the ONE simulation platform.The results reveal that compared with five algorithms,including Epidemic,Spray and Wait,Spray and Wait Ack,MASS,and DPNASW,the FPEU-SAW algorithm can promote the message delivery rate and decrease network latency.(2)A message quality-based cache management strategy(MQ-CMS)is proposed.This strategy analyzes and improves the message discard and relay node selection phases.The message quality is defined by combining the number of message forwarding times,message transmission time,and message remaining lifetime.The message deletion priority is defined by calculating the time spent on message transmission within its lifetime to measure the extent of message diffusion in the network.In addition,the strategy adopts a inter-node buffer state-related message forwarding strategy in selecting relay nodes.In different scenarios,nodes consider the delivery prediction value and node buffer state to decide whether to forward messages to the encounter node.Finally,experiments were conducted using the ONE simulation platform to add this cache management mechanism to six existing algorithms,namely Epidemic,Prophet,Spray and Wait,MASS,DPNASW,and FPEU-SAW.Results of the experiment indicate that this cache management mechanism further promotes the message delivery rate,reduces the hop count of the algorithm,and controls the network load.Moreover,this mechanism demonstrates more significant advantages when the node cache is small.