周惠久论坛-Magnetic Tunneling Junctions Based IoT Data Privacy Protection

作者: 来源:发布时间:2019-04-19


邀请讲座人:张庆瑞 教授

简介:中国台湾大学特聘教授,俄罗斯国际工程院院士,巴黎十三大学荣誉客座教授,国际电机电子工程师学会(IEEE)Fellow,美国物理学会(APS)Fellow。研究领域为Topological Insulator与二维电子系统的自旋传输,磁化的弛缓效应,介观尺寸磁性体的动态翻转机制、多层膜的异向能、巨磁阻来源与应用,磁记录原理及应用。

【报告题目】

Magnetic Tunneling Junctions Based IoT Data Privacy Protection

时间:  4月24日 上午 10:00

地点:  仲英楼 beat365官方网站中文第一会议室

摘要:   


The rapid development of microelectronics industry and spintronics has revolutionized capabilities of Internet of Things (IoT). However, one of the major problem of IoT is that the limitation of the currently available technology cannot simultaneously provided the solution of lower power consumption, high endurance, high security and non-volatile in data process which are critical for IoT application designing. Spin torque transfer magnetic random access memory (STT-MRAM), in addition to a lower power consumption and high endurance in data process, can also use as embedded non-volatile memory in IoT devices. However, security for data and hardware still reserves a tremendous challenge in IoT applications. Motivated by this, we investigate spintronics and propose magnetic tunneling junction (MTJ) based data privacy protection mechanism for data collection from IoT devices equipped with STT-MRAM while satisfying rigorous data privacy guarantee. Analyzing the collected data, we can detect malware activities in IoT devices to achieve the goal of hardware security protection. More precisely, MTJ can generate random events by controlling input voltage and be a hardware random number generator in our system. According to unpredictable period of hardware random numbers, a novel data randomized encoding approach is designed to guarantee data privacy while preserving the feature of population statistics. Through well-designed randomized encoding and the corresponding decoding algorithms in IoT devices equipped with STT-MRAM, our system can not only have rigorous data privacy guarantee but also perform exceptionally in terms of efficient and high-utility malicious behaviors analysis for the collected population of data in protecting hardware security.