Data Center Maintenance Windows Based on Power Consumption Forecasts

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data centers operate at less than full capacity to prevent power outages, leading to underutilized compute resources and potential power degradation during maintenance, which can cause workload interruptions and blackouts.

Innovation Solution

A machine learning model forecasts power consumption to determine maintenance windows where power degradation is minimized, allowing maintenance tasks to be performed during optimal times, reducing the likelihood of workload interruptions and blackouts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data centers operate at less than full capacity to prevent power outages, then reliability is improved, but productivity deteriorates due to underutilized compute resources

Engineering Contradiction:

Improvepower outage preventionVSAvoidcompute resource utilization

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system performs preliminary actions by forecasting power consumption patterns and identifying optimal maintenance windows before actual maintenance occurs. This allows the data center to plan maintenance activities during periods of naturally lower power demand, enabling full capacity operation during normal times while preventing outages during maintenance through advance scheduling.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts maintenance scheduling based on real-time and historical power consumption data. By analyzing patterns and predicting future power usage, the system can flexibly schedule maintenance during optimal windows when power demand is lowest, allowing the data center to operate at full capacity during normal operation while maintaining reliability during maintenance periods.

Inventive Principle:

Principle #15Dynamics

2Productivity

If maintenance tasks are performed during peak power consumption periods, then productivity is maintained, but reliability deteriorates due to increased risk of power degradation and blackouts

Engineering Contradiction:

Improvecompute resource availabilityVSAvoidpower stability during maintenance

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary analysis of power consumption patterns and identifies optimal maintenance windows before scheduling actual maintenance tasks. By predicting when power consumption will be lowest, the system can schedule maintenance during these optimal windows, ensuring that productivity is maintained during normal operation while reliability is protected during maintenance activities.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses feedback from historical and real-time power consumption data to continuously refine maintenance scheduling decisions. By monitoring actual power usage patterns and comparing them with forecasts, the system can adjust maintenance windows to optimize both productivity and reliability, ensuring maintenance occurs when it least impacts power stability.

Inventive Principle:

Principle #23Feedback

3Ease of operation

If maintenance windows are scheduled without considering power consumption patterns, then ease of operation is improved, but reliability deteriorates due to increased risk of workload interruptions

Engineering Contradiction:

Improvemaintenance scheduling simplicityVSAvoidworkload continuity

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system performs preliminary forecasting of power consumption patterns to automatically identify optimal maintenance windows. This preliminary analysis eliminates the need for manual power consumption analysis during scheduling, maintaining ease of operation while significantly improving reliability by ensuring maintenance occurs during periods of lowest power demand and highest workload flexibility.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system serves itself by automatically analyzing power consumption patterns and generating maintenance window recommendations without requiring manual intervention. The self-service approach maintains ease of operation for users while ensuring reliability through data-driven scheduling that automatically selects optimal times based on real-time and historical power usage patterns.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250373502A1Power consumption forecasting and maintenance window determination for maintenance of data centers

Publication Date: 2025.12.04 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Systems, methods, and techniques described herein relate to forecasting power consumption for maintenance of data centers. In an aspect, a maintenance window request for a requested maintenance task with respect to a data center is received. A machine learning (ML) model utilized to generate a window recommendation result, the model trained to determine a period of time where the task's impact on performance of the data center satisfies an impact criterion. The window recommendation result comprising a window indicating a period of time wherein impact of the requested maintenance task satisfies the impact criterion. Performance of the maintenance task is to be performed during the window. In an alternative aspect, the ML model is trained to forecast power consumption of a data center in a period of time. In this alternative, a maintenance window determiner determines the window recommendation result based on a power consumption forecast of the ML model.