Singapore – November 2024 – A leading global cloud services provider has achieved 99.99% operational uptime at its Singapore data center after deploying the ABB SD812F power module series. The upgrade, completed in Q3 2024, addressed critical reliability gaps in the data center’s uninterruptible power supply (UPS) system, which was causing costly outages and risking customer data loss.
The Challenge: UPS Instability Threatens Data Center Reliability
Prior to adopting the SD812F, the data center’s UPS system relied on legacy power modules to provide backup power during grid outages and stabilize voltage for high-density server racks. The old system faced two major issues:
- Frequent UPS Failures: The modules suffered from capacitor degradation, leading to 3–4 UPS failures per year. Each failure lasted 15–20 minutes, disrupting services for 5,000+ customers and costing the provider €250,000 per incident in compensation and recovery.
- Inadequate Scalability: The old modules could only support 5kW per rack, limiting the data center’s ability to add high-power AI and cloud computing servers (which require 8–10kW per rack). This restricted the center’s growth as demand for cloud services surged in Southeast Asia.
- High Cooling Costs: The inefficient modules generated excessive heat, requiring the data center to run cooling systems at full capacity—adding €90,000 to monthly energy bills.
With Singapore’s status as a regional tech hub, the data center needed a power solution that could deliver 24/7 reliability and support future growth. “Uptime is everything in our business—even a 10-minute outage costs us millions in trust and revenue,” said Tan Wei Ling, the data center’s facility manager. “Our old modules were holding us back from serving our customers and scaling our operations.”
ABB SD812F: The Scalable, Reliable Power Backbone
After evaluating four leading power modules, the provider selected the ABB SD812F series for its redundancy, scalability, and low heat generation:
- 99.999% Module Reliability: The SD812F’s rugged design and redundant components (dual capacitors, hot-swappable fans) eliminate single points of failure. Its mean time between failures (MTBF) exceeds 500,000 hours—10x longer than the old modules.
- Scalable Power Output: Each SD812F module delivers 10kW of continuous power, supporting high-density AI and cloud servers. The modular design allows the data center to add modules as needed, without replacing the entire UPS system.
- Low Heat Generation: With 98.8% efficiency, the modules produce 80% less heat than the old system, reducing cooling requirements and cutting energy costs.
- Remote Monitoring & Diagnostics: The SD812F integrates with ABB’s Ability™ Data Center Automation platform, providing real-time alerts for component wear (e.g., capacitor health) and allowing technicians to perform remote troubleshooting.
Results: Uptime Hits 99.99%, Scalability Expands
Four months post-deployment, the Singapore data center has achieved remarkable results:
- Uptime reaches 99.99%: UPS failures dropped to zero, translating to just 4.38 hours of downtime per year—down from 24–32 hours previously.
- Scalability increased by 60%: The data center can now support 10kW per rack, enabling the addition of 1,200 high-power AI servers and opening up €3 million in new cloud service revenue.
- Cooling costs cut by 18%: Reduced heat generation allowed the data center to lower cooling system usage, saving €194,400 annually.
“The SD812F has redefined our data center’s reliability,” Tan said. “We now have the confidence to take on more high-demand customers, knowing our power system won’t let them down. And the scalability means we can grow without ripping and replacing infrastructure.”
David Loh, ABB’s data center solutions director for Asia Pacific, highlighted the SD812F’s fit for modern data centers. “Today’s data centers need power modules that are reliable, scalable, and efficient—all at once,” he said. “The SD812F series checks all those boxes, helping our customers meet the demands of cloud computing and AI while keeping costs in check.”
The cloud provider plans to deploy the SD812F series at its data centers in Tokyo, Seoul, and Sydney by mid-2025.
