Virtualized Cluster Power Management via Workload Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current virtualization-based power-efficiency management methods in data centers fail to maximize power reduction while meeting quality of service requirements, particularly due to slow response to workload fluctuations and inefficient resource allocation among virtual machines.
Innovation Solution
A method and apparatus for power-efficiency management in virtualized cluster systems that detect flow characteristics at regular intervals, generate and implement policies to optimize resource allocation and power consumption, using a front-end and back-end host architecture with virtual machine managers and Hypervisor software.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If virtual machine migration is used for power-efficiency management, then power consumption can be reduced, but the system cannot respond quickly to workload fluctuations and frequent migration increases power consumption
Solution Approach 1:
The patent applies preliminary action by predicting future workload characteristics before they occur. The prediction module analyzes current workload patterns and predicts future resource demands, allowing the system to proactively adjust resource allocation and virtual machine placement before workload changes actually happen. This eliminates the lag inherent in reactive migration approaches and enables the system to respond to workload fluctuations without the delay of traditional migration-based methods.
Solution Approach 2:
The patent implements dynamics by creating a dynamic resource allocation system that continuously adapts to changing workload conditions. Rather than relying on static virtual machine placements or periodic migration, the system dynamically adjusts CPU, memory, and storage resource allocation in real-time based on predicted workload characteristics. This dynamic approach allows the system to optimize power consumption while maintaining responsiveness to rapid workload changes.
2Productivity
If virtual machine manager allocates resources based on current demands, then resource allocation can be adjusted, but the manager cannot sense flow characteristics such as resource demand variation trend affecting optimal allocation decisions
Solution Approach 1:
The patent implements feedback by creating a closed-loop system where workload characteristics are continuously monitored, predicted, and used to inform resource allocation decisions. The prediction module provides feedback about future resource demands to the resource allocation module, enabling continuous optimization. This feedback mechanism allows the virtual machine manager to sense flow characteristics including resource demand variation trends, transforming the system from reactive to proactive in its resource management approach.
Solution Approach 2:
The patent applies preliminary action by predicting future workload characteristics before they occur. The prediction module analyzes current workload patterns and predicts future resource demands, allowing the system to proactively adjust resource allocation and virtual machine placement before workload changes actually happen. This eliminates the lag inherent in reactive migration approaches and enables the system to respond to workload fluctuations without the delay of traditional migration-based methods.
Data Source
AI summary
A method and apparatus for power-efficiency management in a virtualized cluster system. The virtualized cluster system includes a front-end physical host and at least one back-end physical host, and each of the at least one back-end physical host comprises at least one virtual machine and a virtual machine manager. Flow characteristics of the virtualized cluster system are detected at a regular time cycle, a power-efficiency management policy is generated for each of at least one back-end physical host based on the detected flow characteristics, and the power-efficiency management policies are performed. The method can detect the real-time flow characteristics of the virtualized cluster system and make the power-efficiency management policies thereupon to control the power consumption of the system and perform admission control on the whole flow, thereby realizing optimal power saving while meeting the quality of service requirements, so that a virtualized cluster system with high power-efficiency is provided.