Micro Modular Data Center Project for University Research Institute in China
[Abstract]: A Chinese university institute has put Gottogpower’s GT‑DC8000 micro modular data center into use for scientific research computing and digital services.
1. Project Background
As scientific research digitalization continues to accelerate, the institute has become increasingly dependent on high-performance computing platforms, simulation systems, large-scale data storage platforms, and research management information systems. These critical platforms support scientific data processing, computational modeling, experimental analysis, and academic resource management, placing significantly higher demands on data center infrastructure reliability.
However, with the continuous expansion of research activities and IT resources, the original server room environment gradually exposed several operational limitations.
First, the existing power infrastructure lacked sufficient redundancy capacity. During periods of intensive scientific computing workloads, the system struggled to guarantee continuous and stable operation of critical equipment. Any power fluctuation or unexpected outage could interrupt long-running computational tasks, result in experimental data loss, and even delay important research projects.
Second, as the number of high-density servers, GPU computing nodes, and storage systems continued to increase, the thermal load inside the server room rose significantly. The original cooling system was not designed for high-density IT environments, leading to frequent localized hot spots and increasing the long-term operational risk for critical equipment.
At the same time, the traditional server room architecture lacked standardized modular design capability. Future expansion required complex construction work and infrastructure modification, which not only increased deployment cycles but also risked affecting ongoing research operations.
In addition, the existing monitoring system was relatively fragmented. UPS systems, batteries, cooling equipment, and environmental conditions could not be centrally managed through a unified platform, making it difficult for maintenance personnel to identify potential issues in advance and increasing overall operational risk.
To address these challenges, the institute planned to establish a next-generation micro modular data center platform with high reliability, flexible scalability, precise environmental control, and intelligent centralized management capability.
2. Project Requirements
During the planning phase, the customer defined a series of detailed infrastructure requirements focused on long-term research continuity and future development.
The data center had to ensure continuous and stable operation of critical research systems, including high-performance computing clusters, virtualized research platforms, and scientific data service systems. Many research tasks require long-duration uninterrupted computation, meaning even short-term power interruptions could result in failed calculations, interrupted experiments, or significant data loss. Therefore, the power infrastructure needed to achieve extremely high reliability.
The power system also needed to provide high-level redundancy capable of supporting both normal research workloads and periodic high-density computational peaks while maintaining stable output under all operating conditions.
Because the available building space within the institute was limited, the new data center solution had to adopt a highly integrated design capable of consolidating power distribution, cooling, monitoring, and IT deployment into a compact infrastructure architecture while maximizing space utilization efficiency.
At the same time, the customer required strong scalability to support future expansion. As research projects continue to grow, the infrastructure needed to allow flexible addition of IT racks, UPS capacity, and cooling resources without repeated reconstruction or excessive resource waste.
Environmental stability was another major requirement. The server room needed to maintain highly stable temperature and humidity conditions to ensure the long-term safe operation of GPU servers, storage arrays, and network switching equipment while minimizing equipment failure risks caused by environmental fluctuation.
In addition, the institute required a unified intelligent management platform capable of real-time infrastructure monitoring, centralized visualization, alarm notification, and rapid fault response. The goal was to reduce manual inspection workload while improving operational efficiency and infrastructure visibility.
3. Solution
