Continuous Development, Maintenance, and Evolution of Cyber-Physical Systems

点击翻译按钮获取中文摘要

Cyber-Physical systems (CPSs) refer to systems consisting of software and physical components that monitor, control, and coordinate other electro-mechanical units, software, and aspects of the physical world. CPSs encompass a wide range of mission-critical applications that include smart electrical grids, city-wide traffic control, autonomous cars, smart homes, smart farming, the intelligent management of the food supply chain, factory automation systems, and applications that support cost effective health care delivery. Several think tanks expect future CPS applications to be more transformative than the IT revolution of the past three decades. However, these systems pose a set of new challenges related to their modeling, development, deployment and evolution processes. This is not only due to the heavy interactions CPSs software components and devices entail with the physical world, but also due to the practical challenges related to frequent adapt-reuse operations, the need for short release cycles to introduce new features to many users in many platforms with varying requirements, and the need of managing deployment risk in a scale and complexity we have not seen before. Even though the research community has started investigating a number of research areas to address the engineering of CPSs, there is very limited work to date on Development and Operations (DevOps) frameworks to support the continuous development, maintenance, and evolution of such systems. The specific research questions to be addressed in this project are a) how to model and denote dependencies in CPSs (i.e. dependencies between hardware devices, software components, data sources, system stakeholders, third party systems), b) what analytics are required to assist system engineers understand the system's health, identify the system's hot spots, perform root cause analysis, and assess failure risks during maintenance and evolution, and c) devise gated DevOps pipelines to automate build, testing, delivery, deployment, and maintenance processes.