Publikation
Security Validation of VP-based Heterogeneous Systems: A Completeness-driven Perspective
Ece Nur Demirhan Coskun; Muhammad Hassan; Rolf Drechsler
In: Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen (MBMV). ITG/GMM/GI-Workshop "Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen" (MBMV-2023), March 23-24, Freiburg, Germany, 2023.
Zusammenfassung
The widespread use of modern, feature-packed heterogeneous systems has increased the need for robust security measures. A single bug may cause far-reaching and devastating consequences, blocking accessibility of various in-house and third-party Intellectual Properties (IPs) and resulting in an entire system’s failure. In this regard, the concept of Completeness-Driven Development (CDD) provides the promise of early detection of bugs and an accelerated design process. The high-level idea is to use Virtual Prototypes (VPs) at the abstraction of Electronic System Level (ESL) as the starting point for early hardware/software co-design and verification. Going down the abstraction levels, the next abstraction in the design process can only be entered if completeness at the current abstraction level has been achieved.
Completeness refers to checking whether the entire behavior of the design has been verified. However, CDD was introduced for functional verification of digital systems without considering security. In comparison, the modern systems are heterogeneous and security is of utmost importance. In this paper, we look at security validation of VP-based heterogeneous systems from the perspective of CDD. We provide an overview of the current state of security validation techniques and highlight the need for CDD to ensure the security. More concretely, we propose a novel Information Flow Tracking tool – VAST, for complex heterogeneous systems using SystemC-AMS VPs. VAST targets availability of IPs as the threat model. Our experimental results on real-world case-studies show the applicability and scalability of VAST