Preface: Special Issue on Advances in System Software

Abstract

System software, the software that manages the hardware and supports the applications in computing systems, is a core component of IT ecosystem that covers operating system, programming language, compiler, runtime environment, integrated development environment, etc. In industrial competition, system software is the “moat” and “accelerator” of CPU and the “soil” of APPs. The prosperity of academic research on system software in one country often represents its leadership in core software and hardware technologies in the world. Compared with the prosperity of Internet software or smartphone applications, there is still a large gap in the development of system software in China. The research topics on system software usually come from large processor vendors and large application platforms. In the past, there were few research groups in Chinese universities and institutes on system software due to the lack of demand from domestic industrial giants. People have no choice but use the ISAs, micro-architectures, and platforms from international IT giants, e.g. IBM, Intel, and Qualcomm. Therefore, a virtuous cycle of “industrial demand–academic research–production application–new industrial demand” did not exist in China in past decades. The situation has changed in recent years, and system software has become a hot area in both academia and industry for the first time in the history of China. One reason is the rise of domestic chip vendors. The vendors represented by Huawei and Cambricon have taken the lead in the design of smartphone chips and AI chips, and developed their own system software products, and even make them open source, such as openEuler, OpenHarmony, MindSpore, and BANG. Another reason is the expansion of business stack of Internet companies. Internet giants such as Baidu, Tencent, Alibaba, and ByteDance have begun to develop their own chips or invest heavily in the chip industry, and actively joined the game of open source on system software, e.g. Apollo autopilot platform and PaddlePaddle deep learning framework. The cutting-edge demands from chip vendors and Internet giants have ignited the interest of academic research on system software. Many research results have gained international influence, such as the virtualization, IPC/RPC, RDMA, and TEE conducted by Prof. Haibo Chen’s group (IPADS Lab) at Shanghai Jiao Tong University, the graph computing and distributed parallel computing conducted by Prof. Wenguang Chen’s group at Tsinghua University, and the NVM/SSD storage system conducted by Prof. Jiwu Shu’s group at Tsinghua University/Xiamen University.