In recent years, the drive to replace foreign technology with domestic alternatives under China's Xinchuang initiative has continued to accelerate, with home-grown operating systems such as KylinOS and UnionTech UOS being deployed at a faster pace in key sectors including government, finance and energy. Following the policy timeline commonly cited across the industry, 2026 is a core assessment year for Xinchuang, with domestic procurement targets of 85% and 65% for government bodies and state-owned enterprises respectively, while end-2027 is a critical milestone for full-stack replacement. According to industry statistics, domestic operating systems have already passed 85% penetration in government, and reached 62% and 78% in finance and energy respectively. Domestic operating systems as a whole have moved beyond the question of "can it be installed?" — the question enterprises now have to answer is: which system can run stably for three years on my machines, with my software, within my business window?
Drawing on our adaptation experience in government and enterprise Xinchuang projects, this article explains KylinOS and UnionTech UOS adaptation in full — from the ecosystem landscape and full-stack mindset to common pain points, implementation roadmaps and lessons learned.
1. Ecosystem landscape: the "two leading players" of Kylin and UnionTech
First, the supply side. Kylin Software's Galaxy Kylin has ranked first in China's Linux market by share for fifteen consecutive years (CCID Consulting data), with cumulative deployments exceeding 20 million units and more than 210,000 customers. On the desktop side, V11 was released in August 2025; on the server side, the Galaxy Kylin Advanced Server Operating System V10 has obtained Level 4 certification under China's Multi-Level Protection Scheme (the highest security classification for non-classified systems domestically), and V11 SP1 passed Level II safety and reliability assessment in September 2026. Industry data puts Kylin's share of the Xinchuang desktop market at around 70% and its server market share at 80%–90%.
The latest figures published by UnionTech show that its ecosystem now supports more than 10 million compatible items — over 820,000 hardware devices and over 9.2 million software applications — with more than 15,000 ecosystem partners. UnionTech UOS V25, released in April 2026, is built on the Linux 6.6 kernel, features a Windows-friendly interaction design for migrating users, and introduces desktop Agent capabilities. The overall landscape can be summed up as follows: on the desktop, Kylin and UnionTech stand side by side with different emphases; on the server side, Kylin takes the lead; and open-source communities such as openEuler provide the technical foundation for many commercial distributions.
2. Adaptation is a full-stack engineering effort, not just "swapping computers"
Many projects start out treating Xinchuang as simply "switching to domestic PCs", only to discover at acceptance time that the real problems lie elsewhere. At its core, Xinchuang is a wholesale replacement of the IT stack: chips and hardware at the bottom, operating systems, databases and middleware above them, and application software and security components on top. If any single layer is not aligned, the replacement does not hold.
CPU platforms alone present a multi-architecture reality, with domestic architectures such as Phytium, Kunpeng, Loongson, Hygon, Zhaoxin and Shenwei coexisting with the installed x86 base: Phytium and Kunpeng follow the ARM route, Loongson uses its own LoongArch instruction set, and Hygon and Zhaoxin take the licensed-x86 route — each with its own drivers, firmware and compilation toolchains. The same business software may, on different architectures, be a simple "recompile and run" job, or a reengineering effort involving "incompatible JNI interfaces and mismatched dependency library versions". This is a reality we have verified repeatedly in multi-architecture platform adaptation projects.
3. Five high-frequency challenges and how to address them
Based on our project experience, the issues that most often stall delivery can be grouped into five categories:
| Challenge type | Typical symptoms | How to address it |
|---|---|---|
| CPU architecture differences | Incompatible JNI native libraries in Java projects, glibc version traps, C/C++ binaries that cannot run directly | Recompile item by item and sort out dependency chains; prefer dependency versions from each architecture's official repositories; use containers to pin the runtime environment |
| Windows application migration | Legacy business systems tied to IE controls, .NET and proprietary clients | Handle by priority: natively adapt high-frequency core systems; for general office dependencies use Windows compatibility layers such as Kylin KWRE/KARE, replacing them with WPS and domestic browsers; bridge special software with virtual machines or remote application delivery |
| Databases and middleware | Oracle-specific syntax and stored procedures throwing errors on Dameng, KingbaseES and openGauss | Scan for and rewrite SQL dialect differences in advance; run data consistency checks after full plus incremental migration; keep a rollback window open |
| Peripherals, the "last mile" | Printers, document cameras and UKey devices lack native drivers — the system installs but the business cannot run | Verify compatibility model by model during selection; when drivers are missing, compile an adaptation or switch to a substitute model; connect UKey devices through the national cryptographic SKF interface |
| Security and compliance | Numerous and scattered remediation items for MLPS and cryptographic evaluation, prone to repeated rework | Fold requirements such as national cryptographic algorithms, certificate chains and audit logs into the adaptation checklist and address them all in one pass |
4. A six-step adaptation roadmap: an actionable checklist
Regardless of project size, we recommend following a fixed roadmap, with a clear deliverable at every step:
- Asset inventory: catalogue existing hardware models, software lists, peripherals and business controls into a hardware and software asset register, tagging each item's usage frequency and business importance.
- Compatibility assessment: check items one by one against the official Kylin and UnionTech adaptation catalogues, producing a compatibility matrix with red/amber/green grading — green for direct migration, amber for items needing modification, red for items needing a replacement solution.
- Adaptation and modification: for amber and red items, carry out driver compilation, SQL rewrites, national cryptographic changes and control replacement, logging each item in a remediation register.
- Testing and verification: combine functional regression, performance baselines and hands-on peripheral testing, replacing vague claims like "mostly tested" with checklist coverage rates.
- Phased migration: run a small-scale pilot in one real business department, operate in parallel for a period, and roll out fully only after confirming that rollback is possible at any time.
- Operations support: establish a security baseline and a centralised patch source, and connect to the monitoring and alerting system so that every machine can be located and managed.
5. Four lessons for avoiding pitfalls
- Avoid "installed but unused": acceptance criteria cannot stop at "the system is installed" — check the checklist coverage of high-frequency business scenarios; if it is installed but not used, the replacement has not really happened.
- Small scope, long cycle: the pilot should last at least one full business cycle (including month-end and quarter-end peaks); intermittent problems only surface after a long enough run.
- Verify peripherals early: confirm compatibility model by model for printers, scanners and UKey devices before procurement decisions — do not let the "last mile" stall a week before delivery.
- Build documentation and evidence as you go: keep adaptation registers, test records and remediation closure records up to date, as both MLPS assessment and project acceptance will need them.
6. How we can help
Hanshui Info has accumulated complete adaptation and delivery experience across many Xinchuang projects, with a corresponding service for every step of the roadmap above: getting business systems running on KylinOS, UnionTech UOS and domestic CPU, database and browser environments corresponds to our Xinchuang application compatibility and conformance testing service; migrating legacy business systems off Windows / Oracle / IE controls corresponds to our business system Xinchuang migration and modernisation service; for zero-loss database migration and operations support, see domestic database migration and operations support; the "last mile" of printers, document cameras, UKey devices and signature controls is addressed by our dedicated peripheral and business control Xinchuang adaptation service; for multi-architecture server adaptation and performance tuning, see domestic CPU multi-architecture platform adaptation and performance tuning; and for day-to-day operations we provide KylinOS technical support and UnionTech UOS technical support.
If you are planning the pace of your Xinchuang replacement, please call 400-9696-513 to speak with one of our technical consultants — start with an asset inventory and compatibility assessment to establish exactly "what can be migrated directly, what needs modification and what needs replacing", then decide how much of a modification budget to commit.