BMC Reboot Power Cap Maintenance via Threshold Adjustment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In complex systems like servers, existing power control methods often unnecessarily reduce power during a baseboard management controller (BMC) reboot, leading to potential circuit breaker activations and system instability.

Innovation Solution

The system adjusts the overcurrent warning threshold of power supply units before a BMC reboot, using an algorithm that calculates a new threshold based on user power cap settings and PSU efficiency, allowing for maintaining a power cap without excessive power reduction, and enabling or disabling hot spares as necessary to prevent inadvertent circuit breaker activations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the system reduces power level of all components during a controller reboot, then system stability is improved, but power efficiency deteriorates due to unnecessary power reduction

Engineering Contradiction:
Improvesystem stabilityVSAvoidpower efficiency
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system segments power control by component type during BMC reboot. Instead of uniformly reducing power to all components, the method selectively maintains power to non-BMC components (processors, storage, I/O devices) while only reducing power to the BMC itself. This segmentation allows the system to achieve stability during reboot without unnecessarily reducing power to other components, thereby resolving the contradiction between reliability and power efficiency.

Inventive Principle:
Principle #1Segmentation

2Object-affected harmful factors

If the system reduces power to minimum level during BMC reboot, then circuit breaker activation is prevented, but system availability deteriorates

Engineering Contradiction:
Improvecircuit breaker activationVSAvoidsystem availability
Core Design Contradiction:
Object-affected harmful factorsVSProductivity

Solution Approach 1:

The system applies local quality by differentiating power reduction requirements across different component categories. Non-BMC components (processors, storage devices, I/O devices) maintain their operational power levels throughout the reboot process, while only the BMC receives reduced power. This localized approach prevents circuit breaker activation caused by BMC power spikes without forcing the entire system to minimum power levels, thereby maintaining system availability while preventing harmful circuit breaker activations.

Inventive Principle:
Principle #3Local quality

3Ease of operation

If the system uniformly reduces power to all components during reboot, then control simplicity is maintained, but device complexity increases due to unnecessary power management overhead

Engineering Contradiction:
Improvecontrol simplicityVSAvoidpower management overhead
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The system implements dynamic power management by adjusting power levels based on component identity and reboot state. The power management controller dynamically determines which components should maintain power (non-BMC components) and which should reduce power (BMC only) during reboot operations. This dynamic approach automates the complexity of selective power management, maintaining ease of operation through automated decision-making while reducing unnecessary power management overhead compared to uniform power reduction approaches.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS10788876B2System and method to maintain power cap while baseboard management controller reboots
Publication Date: 2020.09.29 DELL PROD LP
  • US10788876B2 patent drawing
  • US10788876B2 patent drawing
  • US10788876B2 patent drawing

AI summary

A system for controlling power to a complex system, comprising a plurality of processors, one or more power supply unit, each power supply unit including an overcurrent warning system and a baseboard management controller coupled to the plurality of processors and the one or more power supply unit, wherein the baseboard management controller is configured to determine whether a power cap control has been enabled and to adjust an overcurrent warning threshold of the overcurrent warning system.