摘要
为了解决高速网络流量分类系统的性能瓶颈问题,提出了一种并行网络流量分类系统负载均衡算法。该算法由静态预分配和动态自适应调整两部分组成,采用基于Hash流表实现负载的静态预分配,根据处理节点的动态反馈对Hash流表进行重映射。通过实验和静态Hash算法以及SHI算法进行了对比,实验结果表明,该算法负载均衡度好、丢包率小、流重映射率低,能够满足并行网络流量分类系统负载均衡的要求。
To deal with the bottleneck of the network traffic classification system for high-speed, an load balancing algorithm for parallel network traffic classification system is presented. The algorithm includes static preassign and dynamic adaption. Static preassign is implemented based on Hash table, when unbalancing adapt to changes based on dynamic feedback from the sensor nodes. At last, the results of comparison experimental with static Hash and SHI demonstrate that this algorithm has a good load balance, low pack lose rate and meets the requirement of load balance of parallel network traffic classification system.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第12期4427-4431,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61163058)
广西自然科学基金项目(2011GXNSFB018076)
关键词
并行
网络流量分类
负载均衡
HASH
自适应
parallel
network traffic classification
load balancing
Hash
self-adaption
