北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2013, Vol. 36 ›› Issue (6): 69-73.doi: 10.13190/j.jbupt.2013.06.015

• 论文 • 上一篇    下一篇

测度属性关系分析的分布式系统异常检测

张乐君, 国林, 张健沛, 杨静, 夏磊   

  1. 哈尔滨工程大学 计算机科学与技术学院, 哈尔滨 150001
  • 收稿日期:2012-12-06 出版日期:2013-12-31 发布日期:2013-10-08
  • 作者简介:张乐君(1979—),男,副教授,博士后,E-mail:zhanglejun@hrbeu.edu.cn.
  • 基金资助:

    国家自然科学基金项目(61100008, 61073041, 61073043);中央高校基本科研业务费专项资金项目(HEUCF061002)

Anomaly Detection for Distributed System Based on Measurement Attributes Analysis

ZHANG Le-jun, GUO Lin, ZHANG Jian-pei, YANG Jing, XIA Lei   

  1. College of Computer Science, Harbin Engineering University, Harbin 150001, China
  • Received:2012-12-06 Online:2013-12-31 Published:2013-10-08

摘要:

针对分布式系统的特点,提出了一种基于网格空间的测度属性关联度计算方法,并以此建立了网络模型来刻画分布式系统测度属性的关系;提出了基于隐马尔可夫模型的测度属性关系网络划分方法;利用最新监测的属性关联度信息,设计了基于测度属性关系网络划分的分布式系统异常检测算法;对多种攻击进行异常检测实验,从而验证了提出方法的有效性.

关键词: 分布式系统, 测度属性, 关系网络划分, 异常检测

Abstract:

The anomaly detection methods based on measurement attributes relationship analysis is proposed. The correlation between the measurement attributes in space grid structure is calculated, and hereby the relationship network model of measurement attributes is established. The network segmentation method is put forward based on hidden Markov model. The anomaly detection algorithm for distributed systems is designed by using the newest monitoring attributes correlation. The anomaly detection experiments demonstrate the effectiveness of the proposed method.

Key words: distribution system, measurement attribute, relationship segmentation, anomaly detection

中图分类号: