摘要
常规动态脱敏方法是一种与关联数据库间单向不可逆映射过程,缺乏脱敏技术的灵活性,限制了技术应用优势的发展与延续。文章针对此问题设计了一种茶杯式动态脱敏技术,通过设置动态脱敏代理模块,利用机器自然语言模式识别算法,提取出用户数据需求中的关键敏感信息,组建临时性脱敏规则进行脱敏处理。其中茶杯–滤网式创新性模式设计,在滤网资源池中增设边缘计算功能模块,便于用户与系统间实时的数据感知服务体验的双向反馈,用户可以直接向资源池申请滤网更换需求,边缘计算后的结论指导脱敏方案的调整,并在评估系统安全性后更换脱敏滤网。在传统脱敏机制上既保留了数据隐私的二次安全性能,也增加了脱敏过程的双向性和灵活性。设计的方法具有一定创新性,在数据安全管理推进方面具有较高的参考价值。
Conventional dynamic desensitization method is a one-way and irreversible mapping process with associated database,which lacks the flexibility of desensitization technology,limits the development and continuation of technology application advantages.In this paper,a tea cup type dynamic desensitization technology is designed to solve this problem.By setting up a dynamic desensitization agent module and using machine natural language pattern recognition algorithm,the key sensitive information in user data requirements is extracted,and temporary desensitization rules are established for desensitization processing.Among them,the tea cup filter innovative mode design adds an edge calculation function module in the filter resource pool to facilitate real-time data perception service experience feedback between users and the system.Users can directly apply to the resource pool for filter screen replacement requirements.The edge calculation results guide the adjustment of desensitization scheme,and after the system security is evaluated,the desensitization filter screen is replaced.Comparing with the traditional desensitization mechanism,it not only retains the secondary security performance of data privacy,but also increases the bidirectional and flexibility of the desensitization process.The method designed in this paper is innovative and has a high reference value in promoting data security management.
作者
韩敏
李永刚
佟雪松
HAN Min;LI Yonggang;TONG Xuesong(State Grid Smart Testing Technology(Beijing)Co.,Ltd.,Beijing 102211,China)
出处
《电力信息与通信技术》
2021年第11期78-84,共7页
Electric Power Information and Communication Technology
关键词
茶杯式
动态脱敏技术
滤网
信息提取
大数据
tea cup
dynamic desensitization technology
filter
information extraction
big data
