Chinese Open-Source Server OS Initiatives Are Starting to Challenge Western Influence

With the rise of web-based applications, services, and more recent Artificial Intelligence (AI) augmentation, the importance of servers, and the Operating Systems (OS) that managed them, has never been more important. Linux has the lion’s share of the server OS market, but Linux comes in different variants. Microsoft (MS) Windows Server has 24% of the market, while Linux 62%. The server Linux OS space has been largely dominated by Red Hat Enterprise Linux (RHEL), Ubuntu Server, and Amazon Linux, among others. openEuler, a Linux variant that had its roots in CentOS, is gaining traction not just in China, but worldwide. The multi-kernel project is now under the wing of the OpenAtom Foundation, a nonprofit organization in China dedicated to advancing open-source software.

In mid-November 2024, ABI Research attended the 5th openEuler Summit, hosted in Beijing. Notable highlights include:

• Jiang Dayong, Chairperson of the openEuler Committee, declared that openEuler has become a robust, open-source development platform that can operate and manage not just digital infrastructure assets such as enterprise servers and cloud installations, but also end-user devices (smartphones, tablets, etc.) and embedded modules. A key philosophy has been to embrace a community-based, resilient development model.
• In those 5 years, openEuler attracted more than 20,000 contributors and over 1,800 organizations, with total downloads exceeding 3.6 million. Three Long-Term Support (LTS) policies have been issued. OS instances have been established in over 150 regions and countries. The latest LTS is the openEuler 24.03 LTS. The next iteration is expected to come out in March 2025. 
• Significant strides were announced to integrate AI support into openEuler. In the 24.03 LTS version, an AI-powered chatbot, EulerCopilot, is embedded in the OS to help generate code, conduct debugging and fault analysis, and assist with general system operation and maintenance.
• At the event, nine organizations (Apusic, China Unicom Digital Tech, eSurfing Cloud, Huawei, Institute of Software Chinese Academy of Sciences, Kylinsec, Peking University and PKU-Changsha Institute for Computing and Digital Economy, Tongji University, and Xiamen University) signed project contribution agreements.

The summit was an impressive display of initiative, talent, and commitment—especially having achieved this level of development within 5 years. There is very much an aspiration to see openEuler not only become a versatile OS platform for a variety of compute platforms, but also to integrate AI into the kernel and supporting applications. Hu Xinwei, Chairperson of the OpenAtom openEuler Technical Committee, stated that “openEuler serves as a converged container stack for both general purpose computing and AI computing and a converged OS to support heterogeneous hardware.” openEuler can, therefore, be installed as a regular Linux distribution or can be deployed in the cloud/embedded devices that run on a variety of Central Processing Unit (CPU) architectures, such as RISC-V, ARM x86 and ARM64, and LoongArch64, among others. During the openEuler & Friends 5.0 session, progress was reported on integrating AI processes into monitoring of upstream software, automated package upgrades, and intelligent reviews to enhance collaboration among community developers and improve the efficiency of software package maintenance.

The OpenAtom Foundation, on behalf of openEuler, has forged technical collaborations with organizations like Apache, Ceph, Intel, Linux, OpenInfra, OpenStack, openHPC, and Spark to address the needs of core industry applications. To enhance the security of the open-source software supply chain, relationships have been established with OpenChain (strives to build trust and efficiency in open-source software management), OpenSSF (aims to sustainably secure the development, maintenance, and consumption of open-source software), and Software Package Data Exchange (SPDX) (provides a consistent approach to represent software information, making it easier to share across organizations and tools).

AI is transforming the compute landscape, especially for the data centers handling the cloud-based AI inference requests and training sessions. The enterprise sector is starting to grasp the opportunities of integrating AI into their back-office and end-customer applications. To support existing application compute, and the latest AI compute loads, the number of servers is expected to grow at a Compound Annual Growth Rate (CAGR) of 15% to 65,000 by 2030. From ABI Research’s own research into next-generation data centers, the number of next-gen data center sites will grow from 5,300 in 2023 to 8,400 in 2030. The associated power capacity that accompanies the deployment of these data centers is expected to grow from 160 Gigawatts (GW) in 2023 to 510 GW in 2030 (see ABI Research’s Next-Generation Data Center Forecasts market data (MD-NGDC-24)). Linux has the largest share (62%) of OS installs on servers, both in the enterprise and data centers, and new Linux OS variants are also taking market share. The openEuler developer community has the digital infrastructure sector in their sights. The openEuler developer community believes that by being able to provide seamless data integration between servers and end-point clients, comprehensive AI integration, as well as collaborations with other open source/open standards-based initiatives will make openEuler an attractive OS for a variety of computing application scenarios.

The momentum building behind openEuler is significant. openEuler is expected to increase its market share, especially outside of China. Concerns about energy consumption are creating waves in the enterprise and data center communities, leading to greater adoption of RISC-based architectures, such as ARM64 and RISC-V. These chipset transitions are leading to OS platform migrations. The development of openEuler is on an accelerated cadence with growing software coding and technology development, not just from Huawei, but also universities and enterprises in China and worldwide. For example, Linaro (United Kingdom) contributes to openEuler software-defined storage and Docker containers that can also run on openEuler.