Rack Power Control Method for Data Center Energy Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data centers, server nodes operate independently, leading to excessive power consumption as they often maintain the same performance state, resulting in inefficient energy usage, as all power supply units are activated regardless of load conditions.
Innovation Solution
A power control method for a rack that calculates the total power consumption of nodes and determines the number of power supply units to be activated, using primary and secondary power supply units in pairs, where primary units provide the duty voltage and secondary units are standby, with different input sources to optimize power distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all power supply units are activated to ensure sufficient power supply, then the power supply reliability is improved, but the power consumption increases excessively
Solution Approach 1:
The patent implements dynamic power supply control by continuously monitoring the total power consumption of all nodes and adjusting the number of activated power supply units accordingly. The RMC calculates the required power supply units based on actual load conditions, turning on or off power supply units dynamically to match the current power demand, thus avoiding excessive power consumption while maintaining sufficient power supply reliability.
Solution Approach 2:
The system changes the operational state parameter of power supply units from a static 'all on' configuration to a dynamic configuration where the number of active units is adjusted based on the total power consumption parameter of the nodes. This parameter change allows the system to optimize the balance between power supply reliability and power consumption efficiency.
2Device complexity
If server nodes operate independently with division of work, then the system complexity is reduced, but the overall power consumption increases due to lack of cooperation
Solution Approach 1:
The patent merges the power management functions of individual server nodes into a centralized rack-level management system. The RMC aggregates power consumption information from all nodes and coordinates the operation of power supply units at the rack level, enabling cooperative power management that reduces total power consumption while maintaining the operational independence of individual nodes.
Solution Approach 2:
The RMC serves multiple functions: it monitors power consumption of individual nodes, calculates total power requirements, determines the optimal number of power supply units to activate, and controls the power supply units. This multi-functional approach enables coordinated power management across the rack without adding significant system complexity.
3Reliability
If primary and secondary power supply units are used in pairs with different input sources, then the system reliability is improved, but the device complexity increases
Solution Approach 1:
The patent segments the power supply units into primary and secondary groups, each with different input sources. The RMC selectively activates either primary or secondary power supply units based on operational needs, creating a modular power supply architecture that enhances reliability through redundancy while managing complexity through organized segmentation.
Solution Approach 2:
The system pre-configures multiple power supply units with different input sources before operation. The RMC has already determined the optimal configuration of primary and secondary power supply units in advance, so when power supply is needed, the system can immediately activate the appropriate units without complex real-time decision-making, thus improving reliability while controlling complexity.
Data Source
AI summary
A power control method of a rack having a plurality of nodes includes the following steps. Power information of each node is received. A total power consumption value of the plurality of nodes according to the power information is calculated. A number of power supply units to be turned on according to the total power consumption value and a maximum supplied power value of a single power supply unit is calculated. At least one primary power supply unit and at least one secondary power supply unit in pairs according to the number of power supply units to be turned on is started. The at least one primary power supply unit provides a duty voltage to the plurality of nodes, and the at least one secondary power supply unit does not provide the duty voltage to the plurality of nodes.