Adaptive Cloud Energy Model Training via Event-Triggered Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for calculating energy consumption in cloud computing environments are inaccurate due to volatile workload patterns, leading to increased resource consumption and performance degradation when constantly updating power consumption models.
Innovation Solution
Adaptive machine learning model training that only updates the power consumption model upon detection of significant events such as hardware changes, software updates, or workload characteristics, reducing resource usage by collecting and transmitting energy consumption data only when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If power consumption models are constantly updated to maintain accuracy with volatile workload patterns, then model accuracy is improved, but computing resource consumption increases and performance degrades
Solution Approach 1:
The patent implements periodic action by updating the power consumption model only at specific intervals or trigger events rather than continuously. The system monitors workload patterns and initiates model retraining only when significant changes are detected, transforming the continuous updating process into periodic discrete events. This reduces the frequency of resource-intensive model training operations while maintaining adequate model accuracy for volatile workload patterns.
2Measurement precision
If power consumption models are constantly updated to maintain accuracy with volatile workload patterns, then model accuracy is improved, but computing resource consumption increases
Solution Approach 1:
The system transforms continuous model updating into periodic updates triggered by workload pattern changes. By implementing event-driven retraining that occurs only when significant workload deviations are detected, the patent reduces the frequency of computationally expensive training operations, thereby lowering energy and computing resource consumption while maintaining model accuracy when it matters most.
Solution Approach 2:
The patent applies parameter changes by dynamically adjusting the training frequency and data collection intensity based on workload volatility parameters. When workload patterns are stable, the system reduces updating frequency; when volatility increases, the system intensifies monitoring and updates the model accordingly. This adaptive parameter adjustment optimizes the balance between model accuracy and resource consumption.
3Use of energy by moving object
If traditional calculation methods are used for energy consumption, then resource consumption is reduced, but accuracy deteriorates due to volatile workload patterns
Solution Approach 1:
The system implements self-service by automatically detecting workload pattern changes and autonomously initiating model retraining without manual intervention. The workload monitoring component continuously analyzes system behavior and triggers model updates only when necessary, allowing the system to self-regulate the balance between accuracy and resource usage based on actual operational conditions.
Data Source
AI summary
Computer-implemented methods for adaptively training a machine learning model for estimating energy consumption in a cloud computing system are provided. Aspects include detecting an update event in the cloud computing system and collecting energy consumption data for the cloud computing system for a time period after the occurrence of the update event. Aspects also include retraining a power consumption model based at least in part on the data collected during the time period and storing the power consumption model in the cloud computing system.