Cloud Data Center UCloud Cutting Energy Consumption for High-Performance Efficiency Case Study

Intel® Data Center Manager improves provisioning efficiency and energy usage to meet high-performance operational standards Business: UCloud is a leading Shanghai-based cloud services provider focused on mobile Internet industries. Challenges • Insights into energy consumption • Real-time cross-platform thermal and power monitoring and temperature • Automated discovery of underutilized servers • Server-level centralized remote access • Server and environmental health analysis • Capacity planning Solution • Detailed discovery session • Intel® Data Center Manager Executive Summary UCloud, a leading Shanghai-based cloud-services provider focused on mobile Internet industries, installed Intel® Data Center Manager (Intel® DCM) in its data center for a test deployment. The company sought to monitor the health and thermals of their data center, and improve its energy efficiency. Leveraging Intel® DCM, UCloud needed granular cross-platform insights to manage server health, load balance more efficiently, analyze and remedy data center cooling issues, and gauge the Power Usage Effectiveness (PUE) of multiple server models. The company deployed Intel® DCM and used its energy, thermal and health monitoring features across a broad spectrum of Original Equipment Manufacturer (OEM) servers and safely raised room temperature in the data center by 3°C. The industry standard of power reduction per degree raised is 3 percent. The initial test deployment indicated that if Intel® DCM was deployed across all of the company’s servers, the annual cooling costs of the data center would be reduced by a stunning 9 percent. Intel® DCM delivers device-level power and thermal data also through a feature that turns servers into wireless sensors. This eliminated the need for intelligent Power Distribution Units (PDUs) and would result in a dramatic reduction in capital expenses across the company operation. Intel® DCM aggregated subcomponent thermal and health data across the devices in the test deployment, allowing IT managers the ability to detect, repair, and redeploy failing devices much more quickly. The cost of downtime per hour is $500,000 USD. The solution also enabled more precise calculations for Mean Time to Repair (MTTR) for devices reaching End-of-Life (EOL) and allowed the safe Case Study | Intel® Data Center Manager Cuts Energy Consumption for High-Performance Operational Efficiency Figure 1. Intel® Data Center Manager Console redeployment of robust units. Intel® DCM reduced downtime an average of 2 hours and would save an average of $1,000,000 USD per incident if deployed across the network. Finally, through Intel® DCM’s health monitoring and utilization capabilities, IT staff optimized server workloads and increased rack density by 10% during the test deployment. The added visibility and operational efficiency enabled by DCM would not only allow UCloud to postpone the purchase of additional server racks, but also delay further expenditures including data center construction and increased space rent. Background UCloud installed Intel® DCM in a test deployment to gain greater insights into the environmental thermal efficiency, server utilization, and component health across its servers. Intel conducted a detailed discovery session of the company server environment to diagnose the specific pain points that its DCM solution would address. The cloud services company provides data center support for mobile Internet industries. UCloud mobile Internet customers require features such as 4k resolution and virtual reality. These extras added to existing robust video streaming demands mean that ultra- reliability and scalability are a must. The bandwidth-intensive requirements combined with the need for high scalability necessitated the optimization of server health for enhanced distribution of