一种随机早期检测技术的机会网络拥塞控制策略  

Random Early Detection Technique of the Opportunistic Network Congestion Control Strategy

作　　者：李向丽[1] 李亚光[1] 

机构地区：[1]郑州大学信息工程学院,郑州450001

出　　处：《小型微型计算机系统》2016年第6期1217-1221,共5页Journal of Chinese Computer Systems

基　　金：河南省教育厅科技研究重点项目(14A520021)资助

摘　　要：在存储空间受限的机会网络中,移动节点为进行消息转发,自身的存储空间被大量占用,如果节点处理缓慢,可能会导致网络拥塞的发生.为了避免拥塞,提出一种基于随机早期检测技术的拥塞控制算法OCCS,该算法首先根据机会网络的特点通过预先设定的队列长度将队列划分为三种类型,然后计算到达队列消息的价值,使节点提前对即将造成拥塞的信息进行处理,尽可能的避免网络拥塞的发生.仿真结果表明,将所提出的拥塞控制策略应用于路由算法中,能够有效提高消息投递率,降低网络负载,但在平均端到端时延上有所增加.因此该策略可以应用于对时延要求不高的星际网络、人口稀少地区网络等外来媒体网络.In the limited storage of the opportunistic network,mobile nodes＇ storage spaces are occupied because of messages forward- ing, which may result in network congestion when the relay node process too slowly. In order to avoid congestion as far as possible, an optimal congestion control strategy （ OCCS ） based on random early detection was proposed which make node deal with information that will cause congestion in advance. Firstly the algorithm divides the length of node queue into three types according to the character- istics of the opportunistic network and then calculates the message value. Numerical results show that the message delivery ratio can be increased efficiently, and the overhead ratio can also be optimized, but the end-to-end latency can be increased slightly. So this strategy can be applied on the exotic media networks, such as interplanetary networks, sparsely populated areas networks etc. which the requirement of delay is not high.

关 键 词：机会网络 拥塞控制 消息价值 随机早期检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]