Software Power Control Using Performance Counters

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Legacy servers lack the ability to monitor their own power consumption, making it economically unfeasible to retrofit them with metering and monitoring, and the cost of new servers can be reduced by not including additional hardware for power monitoring, while power costs are rising in data centers.

Innovation Solution

A software-based power control system that generates a power model for each computing device using performance counters to estimate power consumption and adjust or maintain the device's power state to prevent exceeding a set power cap, allowing for power management without additional hardware.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If legacy servers are retrofitted with metering and monitoring hardware to monitor power consumption, then power consumption monitoring capability is improved, but device complexity and cost increase

Engineering Contradiction:

Improvepower consumption monitoring capabilityVSAvoidhardware complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces physical metering and monitoring hardware with a software-based power control system that uses performance counters and power models to estimate power consumption. This substitutes mechanical/electrical measurement systems with software algorithms, eliminating the need for additional monitoring hardware while achieving power consumption awareness.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces performance counters as an intermediary that indirectly measures power consumption characteristics without requiring direct power monitoring hardware. These counters serve as mediators between the system's operational state and power consumption estimation, enabling monitoring capability through existing software interfaces.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If new servers include additional hardware for power monitoring and measurement, then power consumption monitoring capability is improved, but manufacturing cost increases

Engineering Contradiction:

Improvepower consumption monitoring capabilityVSAvoidmanufacturing cost

Core Design Contradiction:

Measurement precisionVSEase of manufacture

Solution Approach 1:

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If software-based power control is implemented to monitor and adjust power consumption, then power consumption management is improved, but device complexity increases

Engineering Contradiction:

Improvepower consumption management efficiencyVSAvoidsoftware control complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements a self-service power management system where the computing device automatically monitors its own performance counters, estimates its power consumption using local power models, and adjusts its operation to stay within power caps without requiring external monitoring hardware or complex centralized control systems.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent implements feedback mechanisms where performance counters provide real-time operational data, power models estimate current and future power consumption, and the system automatically adjusts its operation based on these estimates to maintain compliance with power caps, creating a closed-loop control system.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS8904209B2Estimating and managing power consumption of computing devices using power models

Publication Date: 2014.12.02 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Power consumption of computing devices are monitored with performance counters and used to generate a power model for each computing device. The power models are used to estimate the power consumption of each computing device based on the performance counters. Each computing device is assigned a power cap, and a software-based power control at each computing device monitors the performance counters, estimates the power consumption using the performance counters and the model, and compares the estimated power consumption with the power cap. Depending on whether the estimated power consumption violates the power cap, the power control may transition the computing device to a lower power state to prevent a violation of the power cap or a higher power state if the computing device is below the power cap.