Congetion Control In Delay Tolerant Networks

Delay tolerant network (DTN) is a new type of network architecture. It is designed for the challenged network environments which are characterized by limited node resources, long delay and intermittent connection. Under the extreme network conditions, one key challenge is how to implement effective congestion control to improve network performance. However, the fundamental assumption in the traditional network model that there exists stable end-to-end path between nodes is not satisfied in DTN context. Therefore the methods based on end-to-end feedback cannot be applied to DTN directly and produce the desired effect. In this dissertation, we make an in-depth investigation on the key technology about DTN congestion control. The primary contributions of this dissertation are concluded as follows.(1) Proposing a congestion level based end-to-end acknowledgement mechanism CL-ACK. It is proved in some practical applications of DTN that the hop-by-hop acknowledgement cannot take place of the end-to-end acknowledgement to provide reliability guarantee. The ACK packets can notify the successful delivery of message and effectively reduce the redundant transmissions. The forwarding mode of ACK greatly affects the overhead and delay of ACKs. CL-ACK uses the ratio of drops over replications to indicate congestion, and then adaptively adjusts the spread manner of ACK packets according to the congestion level. Simulation results show that CL-ACK effectively controls resources consumption, reduces end-to-end delay, and achieves high message delivery rate.(2) Proposing a probabilistic acceptance and drop algorithm PAD which adaptively controls congestion for DTNs. In the intermittently connected DTN environments, message replication is often used to reduce the negative effect impacts on delivery due to network disruption. However, it also may easily incur large resource consumption and finally result in network congestion. In PAD algorithm, the queue length and the input/output rate are combined to detect congestion. Based on the congestion state, each node determines the probability of accepting or dropping message to obtain a good trade-off between high delivery ratio and low overhead. Furthermore, a performance model of the multi-copy routing schemes is constructed, using the birth-death Markov process, to analyze the delivery ratio of message. Theoretical analysis and simulation results show the advantage of PAD.(3) Proposing a novel temporal validity based buffer management scheme TVBBM for Content-centric DTNs. Content-centric network is proposed to address the problem that the host-to-host communication paradigm cannot meet users’ demand for content sharing. In this network context, the data dissemination is driven by the user interests. However, owing to the limited storage capacity of nodes, the efficiency of dissemination is easily affected by network congestion. In practice, the temporal validity of user interest has great significance in message dissemination and drop but is ofen overlooked. In TVBBM, we present the first attempt to apply the validity duration of user interest in message dropping decision, and design a multi-attribute utility to decide the dropping priority of messages from the standpoints of node, message and interest. Simulation results prove that TVBBM can improve the delivery ratio and dissemination speed with low overhead.(4) Proposing a betweenness centrality based congestion-aware routing algorithm BCBCA. In the deterministic DTNs, the network topology changes according to some fixed time schedule. Based on the timetable, a time-varying network can be converted into a static network series, and the optimal routes will be calculated by using the classic algorithms. However most of the existing deterministic routing algorithms lack the ability to adaptively adjust the forwarding routes according to the network state. It causes the network traffic concentrate on some active nodes, which gradually becomes hot spots and finally incurs congestion. According to the evolving graph, BCBCA algorithm adopts an improved Dijkstra algorithm to pick out the top w lowest delay paths, and uses the betweenness centrality as the congestion indicator. Ultimately, BCBCA combines the route delay and betweenness centrality to determine the selective probability of each alternative path. The simulation results show that BCBCA algorithm balances the network load and improves the message delivery ratio with acceptable delay.