Container Power Estimation via Benchmark Regression

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for estimating power consumption and efficiency in cloud data centers focus on hardware and virtual machines, failing to provide granular energy impact assessments for containerized applications, which are essential for Green Service Level Agreements and resource optimization.

Innovation Solution

A method and system for estimating power consumption and efficiency of containerized applications in cloud server systems, using benchmarking and regression analysis to model power consumption based on resource utilization, allowing for the development of application profiles and resource management tools that optimize service provisioning and ensure compliance with Green SLAs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If power consumption estimation methods focus on hardware and virtual machines, then power consumption can be estimated at the infrastructure level, but granular energy impact assessment for containerized applications cannot be provided

Engineering Contradiction:

Improvegranularity of power consumption measurementVSAvoidcomplexity of estimation system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the power consumption measurement by introducing container-specific benchmarking that divides the estimation into base infrastructure power and container-specific power consumption. This segmentation enables granular measurement at the application container level while maintaining system manageability through modular benchmarking procedures.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent uses benchmark applications as intermediaries to measure container power consumption. These benchmark applications run within containers and serve as mediators to capture resource utilization patterns and correlate them with power consumption, enabling indirect but accurate measurement without requiring direct hardware instrumentation of each container.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If traditional virtual machine-based power estimation is used, then infrastructure power management is simplified, but resource optimization for containerized applications is limited

Engineering Contradiction:

Improveresource optimization efficiencyVSAvoidpower consumption

Core Design Contradiction:

ProductivityVSUse of energy by stationary object

Solution Approach 1:

The patent changes the measurement parameters from VM-level metrics to container-level metrics by implementing benchmarking that captures container-specific resource utilization (CPU, memory, I/O). This parameter change enables more precise correlation between container workloads and power consumption, facilitating better resource optimization decisions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent establishes a feedback mechanism where power consumption data and resource utilization data are collected, analyzed, and used to generate actionable insights for resource optimization. This feedback loop enables continuous improvement of energy efficiency by identifying optimization opportunities based on actual container performance and power consumption patterns.

Inventive Principle:

Principle #23Feedback

3Reliability

If Green Service Level Agreements are implemented, then energy efficiency guarantees are provided, but accurate power consumption estimation at application level is required

Engineering Contradiction:

Improveenergy efficiency guaranteeVSAvoidpower consumption estimation accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent performs preliminary benchmarking and regression analysis to establish power consumption models before deploying containers in production. This preliminary action creates baseline data and mathematical models that enable accurate power consumption estimation for Green SLA verification without requiring continuous complex measurements during operation.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces direct physical measurement mechanisms with mathematical modeling and regression analysis. By substituting complex hardware measurement systems with software-based estimation models derived from benchmarking data, the system achieves reliable power consumption estimation suitable for Green SLA guarantees while reducing measurement complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS10432491B2Control device for estimation of power consumption and energy efficiency of application containers

Publication Date: 2019.10.01 INTEL CORP

AI summary

Devices and techniques for estimating power consumption and efficiency of application containers configured to run on a server system, such as a cloud server, are provided. In an example, a method can include creating a benchmark application container, naming the benchmark application container on a host server, collecting power consumption information of the host server, collecting resource utilization information of the benchmark application container using an artificial workload, building a statistical model using the power consumption information and the resource utilization information, and generating a first power model of the host server.