一个基于用户网络行为的访问控制模型  

A Network Behavior-based Access Control Model

作　　者：刘畅[1] 何泾沙[1] LIU Chang;HE Jingsha(School of Software, Beijing University o f Technology, Beijing 100124, China)

机构地区：[1]北京工业大学软件学院,北京100124

出　　处：《信息网络安全》2016年第10期34-39,共6页Netinfo Security

基　　金：国家自然科学基金[61272500];国家高技术研究发展计划(国家863计划)[2015AA017204];北京市自然科学基金[4142008]

摘　　要：针对开放式网络环境中身份认证和行为认证分离这一问题,文章从用户网络行为的规律和特点出发,在传统身份认证的基础上,提出了基于用户行为认证的访问控制模型。文章明确了用户网络行为的定义,随后以用户浏览时间和访问路径作为数据源,将算法得出的计算值与阈值对照,结合时态信息和环境信息对用户行为建模,对用户行为进行验证。对于新用户,基于马尔科夫链对其行为进行检测,对于已存在的用户,根据数据源构建频繁访问有向树,并用均值方差算法对行为进行检测。该模型能够适应用户的行为变迁,并进行自动调节。本框架能有效避免用户账号被恶意盗用等现象,对于提高网络安全具有重要意义。Towards the problem that the separation between identity authentication and behavior authentication in open network environment, we focus on the rules and characteristics o f users9 network behavior. Base on the traditional identity authentication, we put forward the action-basedaccess control model. This paper makes clear the definition o f users* network behavior, take user*sglance time and access path as data source, compared the calculated value gotten from the algorithmw ith threshold, model it w ith temporal inform ation and environment inform ation, then realize user5sbehavior verification. For new users, we check the behavior w ith Markov Chains; for existing users, weestablish the directed tree o f frequent access, using the mean-variance algorithm to detect the behavior.This model can adapt to the changes o f user?s behavior automatically. Moreover, this framework canavoid the phenomenon like malicious misappropriate o f user accounts, which has significant meanings toward network security.

关 键 词：网络行为 行为认证 身份认证 访问控制 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]