Scalable Assurance of Cyber Physical Systems

Cyber-Physical Systems (CPS) are software-reliant systems that interact with the physical world. As such, their kinetic effects frequently have safety-critical consequences. The scientific community has recognized this challenge and created techniques to provide mathematical proof techniques for three main aspects (among many others) of their kinetic effect, namely (i) that the software calculates the correct action on the physical process (e.g., full braking) (ii) at the right time (e.g., 100ms after sensing braking threshold) ensuring the (iii) correct physical effect (e.g., dynamical system reach desired state stopped before hitting the wall). Unfortunately, the application of these techniques become impractical due to scalability issues. In this paper we discuss three key scalability issues: (i) multi-criticality, (ii) artifact size, and (ii) cognitive design overload. These issues are clearly not orthogonal and we will discuss their interactions when we describe them.