Fujian Medical University Union Hospital (FMUUH) is well-known and well-regarded in Fuzhou City, China. The hospital was founded over 100 years ago and has continued to grow into a comprehensive center for healthcare, education, and research. Ranking among China’s top-100 Grade-3, Level-A medical institutions, FMUUH is a facility with expertise in multiple disciplines. For example, FMUUH has a complete range of clinical departments, of which 6 are national key specialties, 10 are provincial key specialties, and 8 are among the first phase of key subject matter practice. The hospital currently has 1,336 beds and receives nearly 1.2 million outpatient and 37,000 inpatient admissions annually — with the number of outpatients expected to increase by 15 percent year after year.
Outdated IT Data Center Faces Bottlenecks
Services at FMUUH have grown quickly in the last decades, and an aging IT data center was unable to keep up with their fast-changing needs. As such, the hospital urgently needed to construct a new IT data center. A requirements survey of the live network found that the hospital faced the following IT service challenges:
• System performance bottlenecks: During peak hours, the Hospital Information System (HIS) was processing on the order of ten thousand outpatient service requests each hour, during which CPU memory usage exceeded 80 percent. The legacy UNIX server-based HIS system was unable to support the hospital’s huge service volumes.
• Long emergency response time: In the event of IT-system emergencies, the hot standby failover took more than 10 minutes to recover, which resulted in a higher risk of service interruptions.
• Slow upgrade of service systems: According to design requirements, the HIS system needed to be flexible and scalable. An additional factor is the possibility for changes to service processes that will arise from continuing reforms within the Chinese medical system. The new HIS system needed to be launched quickly and prepared to develop iteratively. Decades of experience using a closed UNIX server architecture clearly showed the difficulties of developing timely service rollouts.
• Need for future-proof IT platform architecture: The hospital’s overall ICT construction must be future-proof to meet the needs for an evolving system for the next 10 years. This expectation will not only minimize having to upgrade facilities and rebuild services, but will also pave the way for emerging technologies to be added such as cloud computing and Big Data.
A New HIS System Enabled by a Powerful Technical Platform
After fully considering multiple factors and the hospital’s own requirements, FMUUH selected Huawei and B-Soft, an integrated systems vendor, to develop a new HIS system. The new system covers all major aspects within the hospital, including doctors, nurses, pharmacy, medical equipment technologies, finance, medical records, statistics, and admissions and discharge management; and further provides a powerful technical platform for in-depth hospital management.
Huawei’s x86-based KunLun 9016 Mission Critical Server and high-end storage OceanStor 18000 V3 combine to form a local highly reliable solution. The SQL Server is deployed to provide database functions for medical services. Additionally, Web/App services are deployed on x86 virtual clusters.
As part of the underlying architecture for the entire system, two data centers are set up for disaster recovery. The two data centers are physically isolated but logically connected to form a cluster. In this way, services will not be disrupted in the event that one of the data centers suffers a failure. KunLun 16-socket x86 servers are used as the database servers for the underlying HIS system. By using physical partitioning (K-Par) technology, these servers are deployed into both data centers to support multiple database applications such as the HIS, Picture Archiving and Communication System (PACS), and Laboratory Information Management System (LIS). An additional layer of redundancy is provided by the Windows Server Failover Clustering (WSFC), which is set up on the K-Par of the KunLun 9016. The SQL Server uses WSFC services and functions to ensure ‘Always On’ availability.
The OceanStor 18000 high-end storage arrays are configured to work in active-active mode. The database is deployed on the arrays to provide read and write services for KunLun service hosts; in addition, replication links are used for mirroring write I/O in real time. If a cluster node or service fails during system operations all the services carried on the node are automatically transferred to another available node. This failover mechanism ensures the continuity of normal system operations.
Higher Operational Efficiency and Transformation to Smart Healthcare
After the new solution was deployed and went online, FMUUH has dramatically improved the running efficiency of their service systems. Specific benefits include the following:
• Stable and reliable services: By using Huawei’s high-end fault-tolerant solution, the HIS platform achieves architecture-level, device-level, and component-level stability and fault tolerance. The RAS 2.0 technology incorporated into KunLun servers provides a demonstrable reliability of 99.999 percent. This high level of reliability ensures that services will not be interrupted during Operations and Maintenance (O&M), upgrades, or expansion of hospital services; core HIS services will always remain highly available; and reliable systems are designed to not suffer service interruptions or data loss.
• Dramatically optimized system performance: CPU use has decreased from 80 percent to 30 percent, and memory use has dropped from 90 percent to 40 percent. The Solid-State Drives (SSDs) used on KunLun servers operate at a rate of Input/Output Operations Per Second (IOPS) that is 200 times higher and only one-tenth the latency of mechanical, rotating-platter Hard Disk Drives (HDDs). Flash memory technologies have greatly reduced the response latency of the HIS system. The hospital is definitely benefiting from the performance of the enhanced service platform, as doctors have become more efficient in issuing medical orders and the hospital’s workflow for processing financial charges has more than doubled in efficiency. During peak hours, report processing and patient service requests are now handled effortlessly.
• Greatly shortens service rollout time: The KunLun Mission Critical servers platform is underpinned by a complete and mature industry ecosystem. Huawei’s solution decouples applications from hardware to avoid vendor lock-in. As such, the hospital can freely choose Operating Systems (OSs), databases (DBs), and middleware to better adapt to the ongoing innovations in medical services and the accelerated rollouts of new services.
• Reduced Total Cost of Ownership (TCO): One result of the legacy UNIX closed architecture was high application development costs and expensive O&M. In contrast, the open KunLun Mission Critical servers, when combined with high-performance converged storage, greatly improve the system processing performance while having reduced the cost of investment by up to 40 percent.
Huawei and Fujian Medical University Union Hospital have collaborated closely to renovate their IT infrastructure for the hospital’s new data centers. In the future, the two parties expect to gain additional benefits. For example, if and when more KunLun Mission Critical servers are deployed in the hospital, the transition will evolve seamlessly from a ‘same-city active-active’ solution for accelerated service processing to a ‘two cities, three centers’ disaster recovery solution with zero data loss or service downtime — even in extreme situations. All of these enhancements put Fujian Medical University Union Hospital on a fast track to completing a Smart Healthcare transformation journey.