Server Rack Power Management via Node Shutdown
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Server rack systems experience instability and potential data loss when the AC-grid is unstable or internal power supply units fail, leading to inefficient operation or forced shutdowns.
Innovation Solution
A server rack system with a monitoring circuit and a rack management controller that adjusts the operation status of server nodes by lowering power consumption or shutting them down when power supply units fail to output normal voltage, using thresholds to determine the extent of failure and necessary actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the server rack system operates all server nodes at full power, then productivity is maximized, but reliability deteriorates when power supply units fail because the system cannot maintain stable operation
Solution Approach 1:
The monitoring circuit continuously monitors power supply units in advance before failure occurs. When a power supply unit is detected to be abnormal, the system proactively lowers the operation status or shuts down affected server nodes before complete failure happens, preventing system instability and data loss while maintaining overall reliability
Solution Approach 2:
The system dynamically adjusts the operation status of server nodes based on real-time power supply conditions. When power supply units are normal, server nodes operate at full capacity for maximum productivity. When power supply failures are detected, the system dynamically lowers operation status or shuts down nodes to maintain system stability, resolving the contradiction between productivity and reliability
2Reliability
If the monitoring circuit monitors all power supply units continuously, then reliability is improved by detecting failures early, but device complexity increases due to additional monitoring components and control logic
Solution Approach 1:
The monitoring circuit is integrated into the existing power distribution architecture, utilizing available signal lines and control interfaces. The system uses self-service monitoring where power supply units report their status through existing communication protocols, reducing the need for separate dedicated monitoring components and minimizing additional system complexity while maintaining reliable failure detection
Data Source
AI summary
A server rack system including a plurality of power supply units, a monitoring circuit, a rack management controller (RMC), and a plurality of server nodes is provided. The monitoring circuit is for monitoring the power supply units. The RMC is for monitoring the power supply units. When the monitoring circuit and/or the RMC finds that at least one of the power supply units failed to output a normal voltage, an operation status of the server nodes is lowered or at least one of the server nodes is forcibly shut down.