Carbon-Aware Power Manager for Cluster Nodes

Solution Overview

Problem

Existing data center management techniques struggle to balance CO2 emission reduction with maintaining workload scalability and availability, often leading to service disruptions and impact on SLA requirements.

Innovation Solution

Implementing a carbon-aware intelligent power manager that uses machine learning models to predict node criticality and apply power savings measures during periods of low demand for scalability and availability, thereby reducing CO2 emissions without compromising service performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-generated harmful factors

If cluster nodes are shut down or put into power savings modes to reduce CO2 emissions, then carbon emission SLAs are met, but service availability and scalability SLA requirements are impacted

Engineering Contradiction:

ImproveCO2 emissionVSAvoidservice availability

Core Design Contradiction:

Object-generated harmful factorsVSReliability

Solution Approach 1:

The system dynamically adjusts power savings measures based on real-time node criticality assessments. The carbon-aware power manager continuously monitors workload patterns, service dependency graphs, and predicted scalability/availability demands to adaptively apply power savings measures only when safe, thereby reducing CO2 emissions without compromising SLA requirements

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the parameter of power consumption by applying different power savings measures (e.g., sleep modes, frequency throttling, or selective shutdowns) to different nodes based on their criticality. Non-critical nodes experience greater power reductions while critical nodes maintain full operation, achieving emission reduction without service disruption

Inventive Principle:

Principle #35Parameter changes

2Object-generated harmful factors

If cluster nodes are shut down or put into power savings modes to reduce CO2 emissions, then carbon emission SLAs are met, but service scalability SLA requirements are impacted

Engineering Contradiction:

ImproveCO2 emissionVSAvoidservice scalability

Core Design Contradiction:

Object-generated harmful factorsVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary actions by predicting future scalability and availability demands using machine learning models before applying power savings measures. The carbon-aware power manager proactively identifies periods of low demand and applies power savings then, ensuring that nodes can be quickly activated when scalability demands arise without service disruption

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts power savings measures based on real-time node criticality assessments. The carbon-aware power manager continuously monitors workload patterns, service dependency graphs, and predicted scalability/availability demands to adaptively apply power savings measures only when safe, thereby reducing CO2 emissions without compromising SLA requirements

Inventive Principle:

Principle #15Dynamics

3Loss of energy

If power savings measures are applied to reduce CO2 emissions, then energy consumption is reduced, but service disruptions occur due to node shutdown and restart

Engineering Contradiction:

Improveenergy consumptionVSAvoidservice disruption

Core Design Contradiction:

Loss of energyVSReliability

Solution Approach 1:

The system applies local quality by differentiating between critical and non-critical nodes. Power savings measures are applied selectively to non-critical nodes while critical nodes maintain full operation. The carbon-aware power manager uses service dependency graphs to identify which nodes can safely experience power reductions without causing service disruptions

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system performs preliminary actions by predicting future scalability and availability demands using machine learning models before applying power savings measures. The carbon-aware power manager proactively identifies periods of low demand and applies power savings then, ensuring that nodes can be quickly activated when scalability demands arise without service disruption

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250085769A1Carbon-aware intelligent power manager for cluster nodes

Publication Date: 2025.03.13 JUNIPER NETWORKS INC

AI summary

Example power management devices and techniques are described. An example computing device include one or more memories and one or more processors. The one or more processors are configured to determine, based on executing of at least one machine learning model, a measure of node criticality for a node of a cluster. The one or more processors are configured to determine, based on the measure of node criticality, a power savings measure of one or more power savings measures to be applied to the node. The one or more processors are configured to apply the power savings measure to the node.