<p>The buffer resources in a delay-tolerant network (DTN) are extremely limited. When a node uses multiple-copy routing protocols that arbitrarily replicate messages to a relay node, the buffer resources are quickly exhausted, resulting in buffer overflows. To address this issue, we propose a buffer management algorithm that first divides messages into a protected message queue, a semiprotected message queue, and an unprotected message queue, and determines from which queue messages will be dropped based on whether the buffer overflow node is the destination node of the newly received message and the global and local knowledge of the network. In the process of dropping messages, we determine the order of dropping based on the network delivery ratio utility function and the network delay utility function. The simulation results show that the algorithm proposed in this paper has better performance than the drop-oldest, most-forwarded first (MOFO) and weight-based drop (WBD) algorithms.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

A buffer management algorithm for multiple-copy routing protocols in delay tolerant networks

  • Hezhe Wang,
  • Huiqiang Wang

摘要

The buffer resources in a delay-tolerant network (DTN) are extremely limited. When a node uses multiple-copy routing protocols that arbitrarily replicate messages to a relay node, the buffer resources are quickly exhausted, resulting in buffer overflows. To address this issue, we propose a buffer management algorithm that first divides messages into a protected message queue, a semiprotected message queue, and an unprotected message queue, and determines from which queue messages will be dropped based on whether the buffer overflow node is the destination node of the newly received message and the global and local knowledge of the network. In the process of dropping messages, we determine the order of dropping based on the network delivery ratio utility function and the network delay utility function. The simulation results show that the algorithm proposed in this paper has better performance than the drop-oldest, most-forwarded first (MOFO) and weight-based drop (WBD) algorithms.