Machine Learning Energy Consumption Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information handling systems face challenges in accurately estimating process-level energy consumption, with previous power meters often providing estimates that are off by a large percentage, especially under heavy network workloads.
Innovation Solution
The system employs a machine learning model that receives and processes energy data from various components of the information handling system, allowing for the determination of accurate energy consumption levels without relying on assumptions about process, workload, or hardware types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional power meters are used to estimate energy consumption, then the system can provide energy estimates, but the accuracy of energy consumption measurement deteriorates significantly under heavy network workloads
Solution Approach 1:
The patent replaces traditional mechanical power metering systems with a machine learning-based estimation system. The ML model processes multiple input features (CPU usage, memory usage, network I/O, storage I/O, device state) to predict energy consumption, substituting the direct electrical measurement approach with an computational intelligence approach that adapts to varying workload conditions.
Solution Approach 2:
The system changes the parameters used for energy estimation from simple power meter readings to a comprehensive set of system state parameters including CPU usage percentage, memory usage percentage, network I/O throughput, storage I/O throughput, and device state (charging/discharging). This multi-parameter approach enables more accurate energy consumption estimation across different workload scenarios.
2Measurement precision
If the system collects and processes multiple sets of energy data from different components, then the energy estimation accuracy improves, but the system complexity increases
Solution Approach 1:
The machine learning model serves multiple functions: it processes diverse input data from different system components (CPU, memory, network, storage), handles both estimation and prediction tasks, and adapts to various device states. This multi-functionality consolidates what would otherwise require separate specialized systems for each measurement task.
Solution Approach 2:
The patent introduces an intermediary layer (the ML model) that receives raw data from multiple system components and transforms it into accurate energy consumption estimates. This intermediary processes the complex multi-source data internally, presenting a simplified accurate output to the user, thereby managing complexity while maintaining measurement precision.
Data Source
AI summary
An information handling system stores a batch of energy data, and receives different sets of energy data from different components. The system stores the different sets of energy data as the batch of energy data. The system provides the batch of energy data to an input layer of a machine learning model and executes the machine learning model. Based on the execution of the machine learning model, the system determines energy consumption by the different components.