信息通信技术与政策

信息通信技术与政策

信息通信技术与政策 ›› 2021, Vol. 47 ›› Issue (8): 86-91.

上一篇    下一篇

云计算环境用户行为信任评估模型构建研究

Construction method of user behavior trust evaluation model in cloud computing environment

王娜1,赵波2,李昂2,   

  1. 1.中国联合网络通信有限公司智网创新中心,北京100044;
    2.京东集团信息安全部,北京100176
  • 出版日期:2021-08-15 发布日期:2021-08-29
  • 作者简介:
    王娜:中国联合网络通信有限公司智网创新中心安全产品经理,主要研究领域为互联网安全、云安全、5G 安全等
    赵波:京东集团信息安全部安全架构师,主要研究领域为攻防对抗、云安全等
    李昂:京东集团信息安全部解决方案工程师,主要研究领域为数据安全、云安全等

WANG Na, ZHAO Bo, LI Ang,   

  1. 1.Intelligence Network & Innovation Center,China United Network Communications Corporation Limited, Beijing 100044, China;
    2.Information Security Department of JD Group, Beijing 100176, China)
  • Online:2021-08-15 Published:2021-08-29

摘要: 针对传统云计算环境信任模型预测精度过低的问题,提出一种基于深度学习的云计算环境用户行为评估模型构建方法。采用预测鄄残差框架的长短期记忆网络对用户在云计算环境中交互行为集进行时序特征提取,在此基础上结合误差机制修正特征向量,构建云计算环境用户行为信任评估模型。该模型能够实现云计算环境用户访问行为安全的实时监控,解决云计算环境用户实时信任度评估精度过低的问题。根据试验数据表明,云计算环境用户行为信任评估模型构建方法能够有效提升用户可信度实时评估的准确率,可为安全态势分析和推演提供有效的数据支撑。

关键词: 云计算环境, 信任评估, 预测鄄残差框架, 长短期记忆网络

Abstract: Aiming at the problem that the prediction accuracy of traditional trust model in cloud computing environment is too low, this paper proposes a method to build user behavior evaluation model in cloud computing environment based on deep learning.The long-term and short-term memory network of prediction residual framework is used to extract the temporal features of users爷access behavior in the cloud computing environment.On this basis, combined with the error mechanism to modify the feature vector, the trust evaluation model of users 爷behavior in the cloud computing environment is constructed.The model can realize real-time monitoring of user behavior security in cloud computing environment, and solve the problem of low trust evaluation of users in cloud computing environment.Experimental results show that the proposed method can effectively improve the accuracy of real-time evaluation of user credibility,and provide data support for later security situation deduction.

Key words: cloud computing environment, trust assessment, prediction-residual framework, long-term and short-term memory network