SmartX ECP and SKS Empower Core HIS Applications, Accelerating Innovation and Cloud-Native Transformation

As core business systems, the hospital's HIS and EMR are critical to patient efficiency and safety. During its informatics development, the hospital faced challenges such as unrestricted scaling, wasteful resource over-provisioning, complex O&M, long product iteration cycles, and business downtime during upgrades. Simultaneously, seized by the strategic opportunity to build a regional medical center for Neijiang and Zigong, the hospital urgently required a cloud-native digital foundation capable of supporting a microservices-based application architecture to drive high-quality development. Following extensive evaluation and considering the localization requirements of the healthcare industry, Zigong First People's Hospital planned to undergo infrastructure domestic innovation and cloud-native transformation based on SmartX ECP architecture.

Current medical business applications lack adaptation testing for relevant hardware and software platforms, and a complete, mature product ecosystem has yet to form. The hospital determined that transformation should follow a "dual-decoupling" model—addressing applications first, followed by infrastructure. This involves initially using an operating system for application adaptation and upgrades at a low cost, then subsequently utilizing Hyper-Converged Infrastructure (HCI) at the underlying layer to achieve the upgrade and replacement of the entire system.

In collaboration with medical application vendors, the hospital adopted application containerization to achieve decoupling from openEuler. This ensures that each application has its own file system, CPU, memory, and process space, minimizing dependencies on underlying hardware and software and eliminating reliance on specific hardware configurations or operating systems. This approach masks the complexity of heterogeneous underlying infrastructure, reduces dependence on hardware-assisted virtualization, and simplifies compatibility and adaptation between the operating system and software. Meanwhile, by leveraging orchestration engines, the hospital can achieve real-time monitoring and dynamic scaling of application resources, reducing manual intervention and enhancing overall system high availability.

The hospital deployed SMTX OS to pool and service infrastructure resources, including compute, storage, and networking. By adopting Hygon C86 chips during the transformation, the project complexity was effectively reduced. First, the CPU showed high mutual utilization and low actual resource occupancy during virtualization in production. Second, following application containerization, the redundant resource consumption from operating systems, components, and antivirus software was significantly lowered, further reducing CPU load. Finally, by trading core counts for aggregate performance to achieve equivalent output, the hospital successfully completed the adaptation and replacement of its CPUs.

Additionally, the hospital deployed SKS with a single click within the unified cluster management platform. Through a single SmartX ECP management platform, the hospital achieved unified management of both virtualized and containerized business operations. This also enables flat-network interconnectivity between virtual machines and containers, along with unified network security policy management.

Through this project, the hospital has architecturally resolved the adaptation issues between chips, platform software, operating systems, and applications. Currently, part of the core business transformation has been completed, with openEuler 22.03 LTS adopted as the operating system. Simultaneously, cloud-native technology provides autonomous load balancing for high-concurrency services such as HIS. The hospital has deployed 19 virtual machine servers to host the entire business system, including a 3-node high-availability management cluster and several nodes as application instances. During peak hours, application instances scale automatically, and resources are reclaimed once traffic subsides. Meanwhile, backup and reporting instances scale out to achieve elastic expansion and resource sharing, effectively reducing the hospital's hardware investment.