A methodology for transforming a local safety-critical cyber-physical system into a distributed safety-critical solution

Kurzfassung

Cyber-Physical Systems (CPS) integrate computation and physical processes in a complex and dynamic interaction with the real world. CPS are employed in diverse industrial domains with different needs and restrictions, and in particular their application in the safety-critical systems (e.g., cars, ships, or medical devices) is one of the most challenging one. Majority of the safety-critical systems are still deployed as standalone systems that are usually hard to upgrade or extend their functionality while at the same time continuously ensure the system's safety, performance, security, and privacy. The goal of the TRANSACT project is to develop a universal, distributed solution architecture for the transformation of safety-critical cyber-physical systems, from localized standalone systems into safe and secure distributed solutions leveraging edge and cloud computing. Next to the architecture, the TRANSACT project proposes a transition methodology for transforming standalone safety-critical CPS into distributed safety-critical CPS solutions. The transition from an isolated end device to a distributed system is particularly challenging for safety critical systems due to the high demands on the trustworthiness of such a distributed system. Therefore, such a transition demands not only a thorough architectural, design, and validation processes but also a thorough analysis from the business and organization perspective. In the first part of this talk we focus on the TRANSACT Transition Methodology outlining the transformation of the monolithic cyber physical systems to the distributed solution spanning over business, architectural, and organizational areas, taking into account the critical cross-cutting aspects such as safety, performance, security, privacy, regulatory, and certification. Opening the standalone CPS system to the edge/cloud services is very challenging, therefore, the methodology covers also the detailed process helping in the planning and execution of such a transition based on lesson learned from various domain-specific use-cases involved in the TRANSACT project (such as automotive, healthcare, wastewater treatment). The proper requirements handling is important step in the assessing the success of the transition, therefore, in the second part of the talk, we describe generically a stepwise process of defining requirements when transforming an initial CPS into a system with parts distributed over the edge-cloud continuum. This methodology starts by defining what is expected from the overall distributed solution and ends with having the defined requirements for all the components. The methodology leads to early identification of functionality that will be offloaded and early checks if offloading is feasible to all. It supports the creation and integration of all interfaces necessary for offloading but most importantly it guides the definition of requirements and identification of validation and verification needs arising during the transition process. The methodology uses formal abstract system models and simulation of the abstract design variants to establish feasibility of realization of the analyzed requirements. The methodology takes into account that also business and organizational aspects influence whether design variants will lead to a viable solution.