高性能网络隐私泄露行为检测方法和体系结构研究

Personal privacy protection is an essential need of Internet users for future network security system. How to comprehensively, accurately and rapidly detect privacy leaks has become one of the most serious issues of current academic and industrial practitioners. Based on backbone and access networks, this research aims at finding behavior characteristics of privacy leaks and the corresponding high performance detection solution which is specially focused on software algorithm and hardware architecture. The proposed work includes: high performance statistical and content characteristics detecting method, characteristic database of network behavior that is ranked by the level of potential harm, a framework that could integrate all the works mentioned above, etc. Besides, we also plan to design an evaluation system for testing the practicability and effectiveness of our work. Methodologically, we use software and hardware simulators for obtaining a preliminary analysis, build models for obtaining theoretical boundaries and build prototypes on multi-core and FPGA based platforms for obtaining efficiency of deployment. The ultimate goal of this research is to meet the requirements of Internet security for privacy leakage detection methods in the next 5-10 years.

个人隐私保护是人们对未来网络安全设备的基本要求。如何全面、准确、快速的检测隐私泄露成为当前学术界与产业界共同面对的重要课题。本研究立足于骨干网络和接入网，全面研究互联网中隐私泄露行为特征以及相适应的高性能特征检测软件方法和硬件体系结构。主要研究内容包括：基于流统计特征和内容特征的高性能检测方法与体系结构、基于隐私泄露危害等级的网络行为特征库、高性能网络隐私泄露行为检测统一框架等。针对所提出方法，我们拟设计一个隐私保护综合评价体系以评估其实用性和有效性。在性能上，本研究采用软件模拟器和硬件仿真进行初步分析，通过理论模型探讨方案的理论边界，并设计实现基于多核平台的软件原型系统和基于芯片的硬件原型系统，以测试方法和体系结构的实际效果。本研究最终目标为设计满足互联网安全未来5-10年需求的高性能隐私泄露行为检测方法和架构。

个人隐私保护是人们对未来网络安全设备的基本要求。如何全面、准确、快速地检测隐私泄露成为当前学术界与产业界共同面对的重要课题。本研究立足于骨干网络和接入网，全面研究互联网中隐私泄露行为特征以及相适应的高性能特征检测软件方法和硬件体系结构。本研究在高性能网络隐私泄露行为检测框架和特征库、高性能大规模正则表达式匹配硬件加速方法、软硬件结合的专用计算加速方法、高性能URL查找和过滤方法、高性能流表查找方法、高性能恶意网页识别方法等六个方面取得了重要进展。同时，研发了一个高性能网络隐私泄露行为检测系统，通过了第三方机构的测评。本研究已发表学术论文7篇（其中SCI检索论文1篇，EI检索论文6篇），申请专利4项，培养研究生8人（已毕业5人）。相关成果在多个国家重大工程系统中实际应用，取得了非常显著的经济价值和社会效益，获得了中共中央办公厅科学技术二等奖。