Empowering Embedded Systems with Cloud-Native Technologies at CNCC 2023

Forum Chair

Xie Guoqi

Deputy Secretary-General of CCF Embedded Systems Committee, Academic Secretary of YOCSEF Changsha

Director of the Key Laboratory of Embedded and Network Computing in Hunan Province, Professor/Doctoral Supervisor at Hunan University

CCF Senior Member, recipient of the National Youth Talent Program, and Hunan Province Outstanding Youth Fund awardee. He conducts research on key core technologies in embedded computing and talent development, leading the completion of several embedded systems and software, including the vehicle network comprehensive verification platform, 4G/5G time synchronization software for the Internet of Vehicles, V2X scenario verification system, automotive software auxiliary design tools, and the open-source embedded real-time virtual machine ZVM, which have been applied in industrial practice. His research results have been published in conferences and journals such as RTSS, DAC, TPDS, TCAD, TDSC, with four English review papers and two English monographs. He serves on the editorial boards of three SCI journals in the embedded field including JSA, JCSC, and MICPRO.

Virtualization Technology for Embedded Real-Time Operating Systems

As embedded systems become increasingly feature-rich, embedded development presents dual demands for rich functionality and hard real-time performance. The mixed kernel deployment based on operating system virtualization technology is one solution to meet these dual demands. The embedded real-time virtual machine ZVM, led by the Key Laboratory of Embedded Computing at Hunan University, is a virtualization software designed based on the real-time operating system Zephyr, capable of running both Linux and Zephyr as Guest OS on the same hardware platform. Currently, this project has been open-sourced to the openEuler community. This report will introduce the key core technologies for implementing ZVM, actual deployment results, and subsequent feature development and ecological construction.

Yang Pengfei

Associate Professor/Doctoral Supervisor at Xian Electronic Science and Technology University, Special Researcher at the Engineering Research Institute

CCF Senior Member, visiting scholar at the University of Leeds, and director of the Alumni Association for Overseas Students. His research focuses on embedded system architecture and security, leading over ten projects including National Natural Science Foundation, Huiyan projects, military equipment projects, and key industrial projects in Shaanxi Province, publishing over 30 research papers in renowned journals and conferences such as IoTJ, TPDS, TIM, JSA, and Esweek, and awarded the second prize for scientific and technological progress by the China Electronics Society.

Task-Driven Cloud-Edge Collaborative Computing Offloading and Service Deployment

In cloud-edge collaborative systems, edge servers have limited computing and storage capabilities, and different nodes exhibit characteristics such as system heterogeneity and resource diversity. Therefore, dynamic and collaborative task management based on task attributes and the resource status of cloud-edge devices is key to reducing system latency, balancing system load, improving system performance, and maximizing the computing power of cloud-edge collaborative systems. This report will elaborate on diverse task access forms in cloud-edge collaborative scenarios, presenting offloading strategies, service deployment strategies, and QoS-aware offline hybrid strategies for heterogeneous scenarios, along with applications of cloud-edge collaborative computing in engineering fields.

Guan Nan

Associate Professor/Doctoral Supervisor at City University of Hong Kong

CCF Senior Member, obtained his PhD from Uppsala University in Sweden, with research interests including real-time embedded systems, autonomous systems, cyber-physical systems, and the Internet of Things. He received the ACM SIGBED Early Career Researcher Award in 2020 and various best paper awards at prestigious conferences. He has served as chair for RTAS 2022 and program committee chair for several other conferences.

Managing Information Synchronicity in Real-Time Systems

In computing systems, when multiple information streams converge, the information synchronicity constraint requires that data from different streams can only be aggregated when they are generated at closely proximate times. For example, in autonomous vehicles, perception of the external environment often relies on the fusion of measurement data obtained from multiple sensors. If the measurement data from two sensors are generated at significantly different times, their information fusion will not accurately reconstruct the view of the environment. This report will introduce our recent work on how to formally prove the satisfaction of information synchronicity constraints and optimize system design to better meet information synchronicity and classical real-time constraints.

Xu Guizhou

Co-founder of Yihui Information, Core Member of the Operating System Kernel Team

Executive Member of the CCF Embedded Systems Committee, Director of R&D at Beijing Yihui Information Technology Co., Ltd., responsible for national key projects in “Intelligent Robotics” and “Smart Grid,” and the “Microservices Operating System” project in the rail transit field, with 12 authorized invention patents. He leads the R&D of aerospace operating systems, with over 20 satellites currently in orbit and three types of rockets successfully launched. He is responsible for Yihui Information’s university program, establishing joint laboratories with Beihang University, Northwestern Polytechnical University, and Nanjing University of Science and Technology to conduct educational collaborations. The textbook “SylixOS Device Driver Development” co-authored by him was published in 2022.

Task-Critical Cloud-Native Technology

Currently, in task-critical industrial scenarios such as automotive, eVTOL, and drones, the increasing scale of software leads to longer release cycles. Yihui Information researches core technologies such as real-time container technology, service-oriented software architecture, and automated container governance platforms based on the real-time operating system SylixOS and time-sharing partitioned operating system Matrix653, launching a task-critical cloud-native technology system for the embedded field. This aims to achieve a safer and more efficient software design model, decoupling software from hardware, enabling rapid software development and iteration, and assisting critical industries in building industry-leading DevSecOps technology R&D systems, significantly shortening the development cycle of task-critical software, reducing development costs, while improving software reuse rates and system fault tolerance.

Zhou Ting

Senior Engineer at Xian Aviation Industry Computing Institute, Head of New Technology Direction for Embedded Operating Systems

He has long been engaged in cutting-edge exploration, key technological breakthroughs, and application of embedded operating system-related technologies, responsible for several national key pre-research projects, with in-depth research in multi-core deterministic processing, lightweight embedded virtual machines, embedded real-time containers, hybrid critical systems, and information security. In recent years, he has begun researching the application of cloud-native technology in onboard embedded scenarios, achieving a series of research results.

Discussion and Practice of Onboard Cloud-Native Solutions

Cloud-native has become the mainstream direction of cloud computing architecture technology development, a preferred choice for building elastic and scalable application systems, widely applied in IT and industrial fields. In recent years, the aviation sector has begun to explore the application of cloud-native technology in joint combat systems and various aircraft models. Based on existing onboard embedded software stacks, integrating cloud-native technology to build a cloud-native capability system suitable for onboard hybrid critical scenarios, including container runtimes for onboard needs, application orchestration, microservice frameworks, and DevOps toolchains, can drive aircraft equipment to achieve software definition and agile capability generation across generations.