首页 | 本学科首页   官方微博 | 高级检索  
     检索      

Linux的网络转发性能研究
引用本文:郑卫斌,丁会宁,李庆海,张德运.Linux的网络转发性能研究[J].西安交通大学学报,2004,38(2):124-127,135.
作者姓名:郑卫斌  丁会宁  李庆海  张德运
作者单位:1. 西安交通大学电子与信息工程学院,710049,西安
2. 西安交通大学西北网络中心,710049,西安
基金项目:国家“八六三”网络安全管理与测评技术 (863 - 3 0 1 - 0 5- 0 3 ).
摘    要:采用实验与性能"轮廓"相结合的方法,对千兆网络下对称多处理器Linux的NAPI模式转发性能进行研究,发现Linux的多处理器扩展性较差,其网络转发的性能瓶颈点是网络过滤模块,影响Linux性能和扩展性的主要原因是同步开销过大和处理器负载不均衡.为此,采用"中断亲和"的静态调度方法来均衡处理器负载,并降低调度引起的高速缓存命中失败率.改进了网络过滤模块,即减少了两个同步点,降低了同步开销.实验证明所提建议可提高Linux多处理器的性能和扩展性.

关 键 词:Linux  对称多处理器  静态调度
文章编号:0253-987X(2004)02-0124-04

Research on IP Forwarding Performance of Linux
Zheng Weibin,Ding Huining,Li Qinghai,Zhang Deyun.Research on IP Forwarding Performance of Linux[J].Journal of Xi'an Jiaotong University,2004,38(2):124-127,135.
Authors:Zheng Weibin  Ding Huining  Li Qinghai  Zhang Deyun
Institution:Zheng Weibin~1,Ding Huining~2,Li Qinghai~1,Zhang Deyun~1
Abstract:The forwarding performance of the NAPI mode of SMP(symmetric multiprocessor)Linux running on Gigabit networks are studied by experiments and profiles. It is found that the scalability of SMP Linux is poor; Netfilter is the performance bottleneck; and the synchronization overhead and load imbalance among processors are the main causes that influence the performance of Linux. The static scheduling with "Interrupt Affinity" is adopted to balance the load among processors, and to reduce the cache misses due to scheduling. Two locks are eliminated in Netfilter to reduce the synchronization overhead. Experiments show our proposal can improve the performance and scalability of SMP Linux.
Keywords:Linux  symmetric multiprocessor  static scheduling
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号