论文部分内容阅读
在容迟网络(DTN)中节点密度稀疏和节点移动导致网络拓扑结构频繁割裂,消息在传递时无法始终存在一条端到端的连通路径,因此DTN路由算法通常采用存储-携带-转发机制将消息从源节点投递至目的节点。针对上述情况,结合节点间相似性与消息生存时间内节点到达目的节点的概率值,提出一种基于节点相似性的概率路由算法(SBPR),包含消息复制与消息转发2种策略。当持有消息的节点与其他节点相遇时,将消息复制给消息节点相似性较小的节点以提高消息投递率。对于与其相似性较大的邻居节点,如果该邻居节点到达目的节点的概率更大,将消息转发至邻居节点以节省网络资源消耗。实验结果表明,在节点缓存不足的情况下,SBPR在消息投递率、网络负载率及消息丢包数等方面的表现均优于Epidemic,Prophet和First Contact路由算法。
In the DTN, node density is sparse and nodes move frequently resulting in the frequent disconnection of the network topology. When the message is delivered, there can not always be an end-to-end communication path. Therefore, the DTN routing algorithm usually adopts the storage-carry-forward mechanism to change the message from The source node delivers to the destination node. In view of the above situation, a probabilistic routing algorithm based on node similarity (SBPR) is proposed based on the similarity between nodes and the probability of nodes reaching the destination node in the message lifetime, including message replication and message forwarding. When a node holding a message meets another node, the message is copied to a node with a similar similarity between the message nodes to improve message delivery rate. For a neighbor node with high similarity, if the neighbor node has a higher probability of reaching the destination node, the message is forwarded to the neighbor node to save the network resource consumption. Experimental results show that SBPR outperforms Epidemic, Prophet and First Contact routing algorithms in terms of message delivery rate, network load rate, and message packet loss in the case of insufficient buffer.