Skip to content

Amanda-WangXiao/Trusted-Camera-Based-on-TCM

Repository files navigation

Trusted Camera Based on TCM

A Project for National Information Security Competition for College Students.
This work is developed and perfected by both teachers and students. Applying some achievements of Beijing Key Laboratory of Trusted Computing,Beijing University of Technology.The users who concerned about this project should follow the LGPL protocol.

这是一个全国大学生信息安全竞赛的参赛项目。
本作品由师生共同开发、完善。应用了北京工业大学可信计算重点实验室的一些成果,关注本项目的使用者应当遵循LGPL协议。

Introduction

In the current era of advocating "Internet of everything", the Internet of things technology is increasingly applied to our life, and the security of the Internet of things is also gradually valued by everyone. The surveillance camera has a wide range of applications, high requirements for data confidentiality and integrity, which is one of the most noteworthy issues in the Internet of things. Therefore, based on ordinary intelligent camera and trusted computing technology, this work designs and realizes a trusted security camera based on TCM.
Our trusted camera not only has the traditional security protection methods such as password and authentication, but also applies the trusted computing technology to the camera system innovatively: based on the TCM security chip, it guarantees the security of the server, the transmission process and the client based on the trusted chain. Through bottom to top, taking into account software and hardware, and taking measures layer by layer to ensure the data security and controllable behavior of the camera.
At present, trusted computing technology has not been applied in the Internet of things. Therefore, we believe that this work will be an important attempt to apply the domestic trusted computing ecosystem and expand the application scope of trusted computing.

在当前这个倡导“万物互联”的时代,物联网技术越来越多的应用到我们的生活之中,物联网安全也逐渐被大家重视起来。监控摄像头有应用范围广,数据保密性、完整性要求高等特点,是最值得关注的物联网问题之一。因此,本作品以普通的智能摄像头为基础,基于可信计算技术,设计实现了一个基于TCM的可信安全摄像头。
我们的可信摄像头不仅具有密码、验证等传统安全保护方式,更是创新性的将可信计算技术应用到摄像头系统中:基于安全芯片TCM,以可信链为依托全程保障服务端、传输过程以及客户端的安全。通过由底至顶,兼顾软、硬件,层层度量的方式来确保摄像头的数据安全与行为可控。
目前可信计算技术在物联网上尚未有实际落地应用。因此我们相信,本作品将是应用国产可信计算生态圈,扩展可信计算应用范围的一次重要尝试。

NOTE

The basic framework on which the work relies: Trusted Frame【cube-1.3.zip】, TCM simulator【cube-tcm-master.zip】and white list related package 【cube-whitelist.tgz】, must be decompressed under Linux system. If you decompress these packages in other systems firstly and then move these to Linux, there may be linking errors.

作品所依赖的基本框架:可信框架【cube-1.3.zip】、TCM模拟器【cube tcm master.zip】和白名单相关包【cube white list.tgz】,必须在Linux系统下解压缩。如果您先在其他系统中解压缩这些包,然后将它们移到Linux,可能会出现链接错误。

Environment

Ubuntu 18.0 or higher

Acknowledgements

This work won the second prize of National Information Security Competition for College Students.
Thanks to many teachers in the Key Laboratory of Trusted Computing, Beijing University of Technology for their help.
Especially, our teacher Hu Jun provides us with many useful developing tools. Mr Hu's home page is: https://github.com/biparadox.
Developers who are interested in trusted computing can learn from the material contributed by him.

本作品获得了全国大学生信息安全竞赛-作品赛二等奖。
感谢北京工业大学可信计算重点实验室中众多老师的帮助。
特别地,感谢胡俊老师为我们提供了许多可信开发工具,胡老师的主页是:https://github.com/biparadox
有对可信计算感兴趣的开发者可以去学习胡老师贡献的材料。