VM Topology Optimization via Resource Utilization Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data center administrators face challenges in optimizing physical resource allocation to maximize resource utilization and reduce virtual machine sprawl, as load changes constantly, leading to uneven server utilization and underuse.

Innovation Solution

The system predicts future virtual machine behaviors by modeling past resource utilization data and preconfigures the virtual machine topology to optimize resource allocation, using a data center optimization system that includes a communications engine, predictions engine, and optimization engine to adjust VM topology based on predicted utilization and physical topology.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If virtual machines are spawned on demand to handle increasing load, then system responsiveness is improved, but virtual machine sprawl increases and resource distribution becomes uneven

Engineering Contradiction:

Improvesystem responsivenessVSAvoidvirtual machine sprawl

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by predicting future resource utilization patterns and proactively migrating virtual machines before load changes occur. The prediction engine analyzes historical data to forecast future states, and the optimization engine pre-configures the virtual machine topology accordingly, preventing over-provisioning and sprawl before it happens.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements continuous feedback loops where resource utilization measurements are collected, predictions are generated based on historical data, optimizations are applied, and results are measured. This closed-loop feedback enables the system to learn from past behaviors and continuously improve virtual machine allocation, balancing responsiveness with resource efficiency.

Inventive Principle:

Principle #23Feedback

2Reliability

If virtual machines are kept running to maintain service availability, then reliability is improved, but resource underutilization increases

Engineering Contradiction:

Improveservice availabilityVSAvoidresource underutilization

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The system makes the virtual machine topology dynamic by continuously predicting future resource needs and automatically migrating or consolidating virtual machines based on forecasted utilization. This dynamic adaptation allows the system to maintain service availability when needed while consolidating resources during low-utilization periods, eliminating the need to keep virtual machines statically running.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes operational parameters by adjusting virtual machine placement and configuration based on predicted resource utilization patterns. By analyzing historical measurement data and forecasting future states, the system optimizes parameters such as virtual machine location, resource allocation, and topology configuration to balance availability with energy efficiency.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If resource allocation is optimized in real-time based on current load, then resource utilization is improved, but system complexity and response time increase

Engineering Contradiction:

Improveresource utilizationVSAvoidoptimization response time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary optimization actions by predicting future resource utilization patterns and pre-configuring virtual machine topology before actual load changes occur. This proactive approach allows the system to achieve optimal resource utilization without the time delay associated with reactive real-time optimization, as the system is already prepared for anticipated future states.

Inventive Principle:

Principle #10Preliminary action

4Measurement precision

If measurements are collected continuously to track resource utilization, then measurement precision is improved, but system overhead increases

Engineering Contradiction:

Improveresource utilization accuracyVSAvoidmeasurement overhead

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system applies partial measurement action by collecting resource utilization measurements at strategically selected intervals rather than continuously. The prediction engine uses these periodic measurements combined with historical data to accurately forecast future utilization patterns, achieving sufficient measurement precision while significantly reducing the overhead associated with continuous monitoring.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10365940B2Feedback system for optimizing the allocation of resources in a data center

Publication Date: 2019.07.30 CLOUBRAIN

AI summary

To improve resource utilization and reduce the virtual machine sprawl in a data center, resource utilization is predicted based on previously measured utilizations, and then, using the predicted utilizations, optimizing the allocation of the computing resources among the virtual machines in the data center. In operation, measurements related to resource utilization by different virtual machines executing in a data center are collected at regular intervals. At each interval, an optimization system predicts virtual machine resource utilizations based on previously collected measurements and previously-generated virtual machine modelers. Based on the utilization predictions as well as the physical topology of the data center, the optimization system identifies different optimizations to the virtual machine topology for the next interval.