基于云平台追责的日志过滤方法研究
首发时间:2018-08-02
摘要:云计算已经成为主流的互联网计算模式。现有的云平台都有着庞大的组织,云平台计算节点数目庞大且分布广泛,节点及组件模块的日志数据量庞大且分散,同一个操作的相关日志可能位于不同的虚拟机实例上,甚至是不同的物理机上,这给基于云平台的日志分析带来了巨大的困难。本文提出了一种量化实例、计算节点和组件(统称为实体)之间相关性的模型。在此基础上,本文提出了一种基于日志相似性递归过滤和实体间相关性的日志过滤方法。经过对比实验表明,此方法比以往的方法有更好的日志过滤效果。
For information in English, please click here
Research on Log Filtering Method Based on Cloud Platform Responsibility
Abstract:Cloud computing has become the mainstream Internet computing model.Existing cloud platforms have huge organizations,the number of compute nodes on the cloud platform is large and widely distributed,the log data of nodes and component modules is large and scattered.Logs related to the same operation may be located on different virtual machine instances or even on different physical machines, which brings great difficulties to cloud-based log analysis.This paper proposes a model for quantifying the correlation between instances, compute nodes, and components (collectively referred to as entities).On this basis, this paper proposes a log filtering method based on log similarity recursive filtering and inter-entity correlation.Comparative experiments show that this method has better log filtering effect than the previous method.
Keywords: cloud platform log correlation log filtering entity
基金:
引用
No.****
同行评议
共计0人参与
勘误表
基于云平台追责的日志过滤方法研究
评论
全部评论0/1000