Dynamic Processor Power Management for Cloud Infrastructure

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current cloud computing infrastructure faces inflexibility in scaling and descaling computing resources, leading to poor resource utilization and increased costs due to fixed hardware configurations and limitations in adjusting processor clock speeds and voltages based on workload demands, resulting in inefficient power management and potential penalties from utility companies for erratic power usage.

Innovation Solution

A power management system for disaggregated computing environments that dynamically adjusts processor clock speeds and voltages based on workload SLAs, utilizing a power management module to predict resource usage and allocate power efficiently across workloads, ensuring maximum utilization of contracted power while maintaining performance and throughput within allocated power ranges.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If fixed hardware configurations are used in cloud computing infrastructure, then device stability and reliability are improved, but adaptability and resource utilization deteriorate

Engineering Contradiction:
Improvesystem stabilityVSAvoidresource scaling flexibility
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic power management by continuously monitoring workload demands and adjusting processor clock speeds and voltages in real-time. The system transitions from static fixed configurations to dynamic adjustable parameters, allowing the infrastructure to adapt to changing computational requirements while maintaining system stability through controlled adjustments.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes physical parameters (clock speed, voltage) of processors based on workload demands. By modifying these parameters dynamically, the system achieves both reliability (through stable operation at appropriate levels) and adaptability (through parameter adjustment to match computational needs).

Inventive Principle:
Principle #35Parameter changes

2Reliability

If processor clock speeds and voltages are kept constant, then device reliability is improved, but energy efficiency and productivity deteriorate

Engineering Contradiction:
Improveprocessor stabilityVSAvoidworkload processing efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent makes processor clock speeds and voltages dynamic rather than constant. The system monitors workload intensity and adjusts these parameters accordingly, enabling the processor to operate at optimal performance levels for each workload while maintaining stability through controlled adjustment mechanisms.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system actively changes processor parameters (clock speed, voltage) based on real-time workload assessment. This allows the processor to deliver high productivity when workloads require it while maintaining reliability through stable operation at appropriate parameter levels, resolving the contradiction between constant parameters and efficiency.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If high processor utilization is maintained continuously, then productivity is improved, but energy consumption and cost increase

Engineering Contradiction:
Improvecomputational throughputVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent changes processor operating parameters (clock speed, voltage) dynamically based on actual workload demands. During high-utilization periods, parameters are increased to maintain productivity; during low-utilization periods, parameters are reduced to decrease energy consumption, thus resolving the contradiction between continuous high productivity and excessive energy use.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system implements periodic monitoring and adjustment of processor parameters based on workload patterns. This periodic adaptation allows the system to maintain high productivity when needed while reducing energy consumption during periods of lower demand, achieving a balance between throughput and power usage.

Inventive Principle:
Principle #19Periodic action

4Productivity

If power allocation is increased to ensure maximum performance, then productivity is improved, but energy cost and loss increase

Engineering Contradiction:
Improvesystem throughputVSAvoidunnecessary power consumption
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent implements dynamic parameter adjustment where processor clock speeds and voltages are changed based on actual workload requirements. This ensures power allocation matches actual productivity needs, avoiding unnecessary energy consumption when full performance is not required while maintaining high throughput when workloads demand it.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system employs feedback mechanisms to monitor workload demands and adjust power allocation accordingly. This closed-loop control ensures that power is allocated efficiently - increasing it when productivity is needed and reducing it when performance requirements are lower, thereby minimizing energy loss while maintaining optimal throughput.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS10545560B2Power management and utilization detection of computing components
Publication Date: 2020.01.28 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US10545560B2 patent drawing
  • US10545560B2 patent drawing
  • US10545560B2 patent drawing

AI summary

For power management in a computing system, component utilization is dynamically managed within the computing system according to a calculated aggregate energy consumed by each one of a set of processors. Each of a plurality of energy factors are measured individually between each one of the set of processors to accumulate the calculated aggregate energy in real time.