EMBOSOM - Emigrating Embedded Software Security into Modern Emerging Hardware Paradigms

为了保护软件免受漏洞的利用，最重要的是要有足够的保护机制。这在嵌入式领域尤其重要，因为此类漏洞可能在现实世界中造成危害。然而，先前的工作已经表明，通用的软件保护机制在这个领域是不可用的。虽然最先进的保护机制主要在软件中实现，但一个新兴的研究方向是直接在硬件中实现这些技术。不幸的是，该领域的现有工作针对具有不同威胁模型的非约束传统设备。此外，这种硬件保护机制的现有工作对嵌入式领域的基本特征（如芯片尺寸、性能和生产成本）产生了重大影响。我们希望开发一种新的基于硬件的安全机制，专门针对嵌入式领域的威胁模型进行定制，并利用新兴技术（例如基于忆阻器的内存计算）来减少对性能，功耗和芯片尺寸的影响。该提案属于研究领域3（安全组合和集成）和通过硬件支持的呼叫矩阵的数据分离和隔离（IG2）进行安全处理的跨学科小组。

In order to protect software against the exploitation of vulnerabilities, it is paramount to have sufficient protection mechanisms in place. This is especially important in the embedded domain, where such vulnerabilities can cause harm in the real world. However, prior work has shown that common software protection mechanisms are not available in this domain. While state-of-the-art protection mechanisms are mainly implemented in software, an emerging research direction is implementing such techniques directly in the hardware. Unfortunately, existing work in this domain targets non-constrained conventional devices which have a different threat model. Furthermore, existing work on such hardware protection mechanisms imposes a significant impact on essential characteristics of the embedded domain such as chip size, performance, and production costs. We want to develop a novel hardware-based security mechanism which is specifically tailored to the threat model of the embedded domain and leverages emerging technologies (e.g. memristor-based in-memory computing) to reduce the impact on performance, power consumption, and chip size. The proposal belongs to research area 3 (Secure Composition and Integration) and the interdisciplinary group on secure processing via hardware-supported data separation and isolation (IG2) of the call's matrix.