Performance Analysis And Optimization Of Delay Tolerant Network Based On Queuing Model

The exploration of the unknown space is becoming more and more frequent.Among them,deep space exploration is an important project for human beings to expand their cognition.Communication system plays an important role in deep space exploration.Compared with the ground communication environment,in deep space,the communication environment is more complex and harsh.That the end to end path does not exist,the lower SNR,the frequent link connection disruption and so on,all of which have brought great challenges to the communication system.The DTN network has been proposed in order to solve the communication problem in the extreme environment.Different from the traditional TCP/IP network,DTN uses a special transmission strategy-hop by hop custody transmission.In the process of data transmission,the data will be stored in the storage space of the DTN node,and it will be deleted from the storage space when it is successfully transferred to the next hop.Due to different data transmission strategy,the architecture of DTN network is different from traditional network protocol architecture.The main difference is that the DTN network joins the bundle protocol under the application layer.Custody transfer function is mainly achieved by the bundle protocol layer.DTN networks have been deployed and used in deep space networks.It is necessary to improve the transmission performance of DTN network.Custody transfer provides the ability to overcome extreme environmental challenges for DTN,but also brings other performance challenges.Analyzing and optimizing the performance of DTN architecture is the key to improve the communication ability of DTN network.The Licklider Transmission Protocol(LTP)is proposed to further improve the efficiency of data transmission over links with long delay and frequent interruptions,which operates under the bundle layer.Bundle protocol and LTP protocol form the DTN architecture-BP/LTP architecture,which is often used in the space network.In this first research point,we focus on such kind of BP/LTP based DTN architecture and analyze its performance from the viewpoint of queuing theory.The distinguishing feature of our work is that we take account of random link interruptions in our analysis.In DTNs,some link interruptions occur inevitably and randomly,having a significant impact on the performance of data delivery.Considering the service time of bundle obeys the general distribution,we model the queue at the bundle layer as an M/G/I queue.Firstly,we derive the probability distribution of the bundle service time under random link interruptions.Based on this,the stability condition of the bundle queue is given,which imposes constraints on the bundle arrival process and the link interruption model.Furthermore,we analyze the impact of random link interruptions on the bundle service process and derive the average delay of the bundle queue.Finally,numerical analysis and simulations are conducted to validate our theoretical analysis.The comparison results show that our model better reflects the impact of random interrupt on the DTN architecture.In the second research point,for the following reasons:1)the storage space in the space network is expensive and limited.2)In space network,large amounts of images and other data need to be transmitted.The way of DTN storage and transmission will retain large amounts of data in the nodes,and the storage space will be greatly consumed.3)For a relay node in a special position,a large amount of data will pass through the relay node.Storing these data results in greater storage pressure.Based on the above reasons,it is necessary to optimize the relay nodes in the BP/LTP network.According to the data processing method of BP/LTP architecture,we build batch queuing model,and analyze the relationship between LTP block,LTP segment size and the average length of bundle queue.Then,we design a joint optimization algorithm to optimize the related variables to reduce the average queue length of bundle,and reduce the consumption of storage queue for storage space.Finally,the simulation results show the effectiveness of our algorithm.