服务器多冲突风暴节点的检测器设计与实现

发布时间：2018-04-21 05:35

  本文选题：服务器 + 冲突　； 参考：《现代电子技术》2017年18期

【摘要】：随着服务器系统节点数量和用户任务的不断增多,使得风暴节点出现冲突问题。在此背景下,通过分析服务器系统的节点连接架构,将检测器部署在其通信平台上,对用于检测服务器多冲突风暴节点的MooseFS检测器进行设计与实现。当风暴节点中存在较多冲突时,MooseFS检测器采用动态冗余分布策略与硬件平台协同检测,冲突不多时只使用硬件进行检测。动态冗余分布策略选用冗余度度量节点任务开销,给出风暴节点子任务、带宽和中央处理器运行时间的可用度函数,检验冗余度是否可靠。硬件平台通过逻辑门处理器对子任务、带宽、中央处理器运行时间以及存储容量的负载率进行度量。实验结果表明,所设计的MooseFS检测器拥有很强的资源控制能力。
[Abstract]:With the increasing number of server nodes and user tasks, storm nodes conflict. Under this background, the MooseFS detector used to detect multi-conflict storm nodes is designed and implemented by analyzing the node connection architecture of the server system and deploying the detector on its communication platform. When there are more conflicts in the storm nodes, the dynamic redundancy distribution strategy and hardware platform are used to detect the dynamic redundancy distribution strategy, and only hardware is used to detect the conflicts. The dynamic redundancy distribution strategy uses redundancy to measure node task overhead. The availability functions of storm node subtask bandwidth and CPU running time are given to verify the reliability of redundancy. The hardware platform measures sub-tasks, bandwidth, CPU runtime and load rate of storage capacity through logic gate processors. Experimental results show that the designed MooseFS detector has a strong resource control ability.
【作者单位】： 北京信息职业技术学院计算机与通信工程学院;中国人民公安大学;
【基金】：国家自然科学基金(71203229) 北京市教委科技计划面上科研项目(KM201310857001)
【分类号】：TP368.5
，

本文编号：1781132