CAREER: Securing Mobile Devices by Hardening their System Software

现代移动设备，如智能手机、平板电脑和可穿戴设备，是越来越多的安全攻击的目标。此外，根据谷歌的说法，针对（基于android的）移动设备的攻击越来越多，目标是操作系统内核。例如，2016年有44%的攻击是针对内核的，而2015年和2014年分别为9%和4%。不幸的是，当前的移动操作系统庞大、复杂且充满漏洞；因此，他们往往成为这些攻击的受害者。该项目试图通过一套互补和全面的方法来改善移动设备的安全现状，这些方法可以减轻操作系统易受攻击的问题。本项目增强了移动设备的安全性。随着移动设备在当今世界发挥越来越重要的作用，这个项目将对全球社会和经济产生重大影响。该项目还将通过外展和辅导活动培训学生进行系统软件编程。该项目的目标是采用三种方法来增强移动设备的安全性：(i)一个安全监控器，在操作系统受损的情况下提供重要的安全和隐私保障；（ii）一种新颖的移动专用工具，用于发现（然后修复）移动操作系统漏洞；以及（iii）一个审查层，有效地保护操作系统界面免受恶意应用程序的侵害。该项目将分三个研究重点进行，每个研究重点将解决上述方法之一的挑战。第一个研究重点是为移动设备构建一个值得信赖和可扩展的安全监视器。这个要点的关键思想是使用安全域，这些域是独立的域，每个域托管不同的安全服务，提供唯一的安全或隐私保证。第二个研究重点是研究在移动操作系统中发现漏洞的技术。特别地，它研究了将现有的动态分析技术应用于移动操作系统代码库的解决方案，以发现和修补其漏洞。大多数此类技术仅适用于在虚拟机（VM）中运行的软件。因此，这项研究的关键思想是要么自动移植移动操作系统代码在虚拟机中运行，要么在设备本身上运行这些技术。第三个研究重点是研究保护操作系统界面免受恶意应用程序侵害的方法。它通过对系统调用和其他操作系统应用程序编程接口调用合并一组广泛的安全检查来限制应用程序的攻击向量。该推力由两个任务组成，包括能够安全有效地执行审查层安全检查的进程内屏蔽空间，以及自动生成此类安全检查的解决方案。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Modern mobile devices, such as smartphones, tablets, and wearables, are targets of an increasing number of security attacks. Moreover, according to Google, an increasing number of attacks on (Android-based) mobile devices are targeting the operating system kernel. For example, 44% of attacks in 2016 targeted the kernel versus 9% and 4% respectively in 2015 and 2014. Unfortunately, contemporary mobile operating systems are large, complex, and full of vulnerabilities; hence they fall victims to these attacks more often than not. This project attempts at improving the state of the art in security of mobile devices through a complementary and comprehensive set of approaches that can alleviate the problem of operating system vulnerability to attacks. This project enhances the security of mobile devices. As mobile devices play an increasingly important role in today's world, this project will have a high impact on global societies and economies. The project will also train students in systems software programming through outreach and tutorial activities. The project targets three approaches to enhancing the security of mobile devices: (i) a security monitor to provide important security and privacy guarantees despite a compromised operating system, (ii) novel and mobile-specific tools to find (and then fix) mobile operating system vulnerabilities, and (iii) a vetting layer to efficiently safeguard the operating system interface against malicious applications. This project will be conducted in three research thrusts, each of which will address the challenges of one of the aforementioned approaches. The first research thrust builds a trustworthy and extensible security monitor for mobile devices. The key idea in this thrust is the use of security domains, which are isolated domains each hosting a different security service providing a unique security or privacy guarantee. The second research thrust investigates techniques to find vulnerabilities in the mobile operating system. In particular, it investigates solutions to apply existing dynamic analysis techniques to mobile operating system codebase in order to find and patch its vulnerabilities. Most such techniques are only applicable to software running within a virtual machine (VM). Hence, the key idea in this research thrust is to either automatically port mobile operating system code to run within a virtual machine or run these techniques on the device itself. The third research thrust investigates methods to safeguard the operating system interface against malicious applications. It does so by incorporating an extensive set of security checks on system calls and other operating system application programming interface calls to limit the application's attack vector. This thrust consists of two tasks including an in-process shield space capable of securely and efficiently executing the vetting layer's security checks as well as solutions to automatically generate such security checks.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.