| 多源日志存储和分析在网络事件关联中的实际应用 |
| |
| 引用本文： | 孙毅,赵泽宇,闫华.多源日志存储和分析在网络事件关联中的实际应用[J].中国海洋大学学报(自然科学版),2008(Z1). |
| |
| 作者姓名： | 孙毅 赵泽宇 闫华 |
| |
| 作者单位： | 复旦大学网络中心 |
| |
| 摘 要： | 网络管理要对网络事件有全面认知。目前网络管理的常用手段主要依靠对所有网络设备和网断的系统信息检测和读取信息,而网络信息庞杂且时刻变化,对信息存储和解读提出了很高的要求和挑战。为了解决上述问题,基于Loglog算法提出了1种全新的网络运行日志存储和恢复算法,并且结合实际运行环境给出1个应用模型,扩展网络管理和维护的时间跨度和空间维度。同时,算法结合网络日志的特点加之最大熵方法的过滤器,并在网络认知关联上有实际的应用,对网络信息挖掘有理论联系实际的意义。
|
| 关 键 词： | 网络事件 Loglog算法 最大熵 基数 网络认知 |
Network Events Relationships Base on Multi Sources Log's Storage and Analysis |
| |
| SUN Yi,ZHAO Ze-Yu,YAN Hua.Network Events Relationships Base on Multi Sources Log''s Storage and Analysis[J].Periodical of Ocean University of China,2008(Z1). |
| |
| Authors: | SUN Yi ZHAO Ze-Yu YAN Hua |
| |
| Institution: | SUN Yi,ZHAO Ze-Yu,YAN Hua(Information Center,Fudan University,Shanghai 200433,China) |
| |
| Abstract: | Considering the characteristic of Network events, maximum entropy model was applied to analyse a huge log file.The LogLog algorithm makes it possible to estimate in a single pass and within a few percents the number of different words in all of network flows.Maximum entropy model is applied to netflow analyse,improved definitions of feature function are given to construct better feature vectors for network events.Furthermore,the framework of log analys is system based on maximum entropy model is constructed... |
| |
| Keywords: | network event loglog counting maximum entropy cardinalities network cognize |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
