Virtual Machine Resource Allocation via Load Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud computing environments, existing technologies face challenges in maintaining performance stability while optimizing resource utilization due to fluctuating loads, leading to high calculation processing demands.
Innovation Solution
A method is introduced that involves classifying load fluctuations into groups, predicting future loads, and allocating physical computing resources based on these predictions to create a schedule for optimal resource allocation, reducing calculation processing and ensuring performance stability and resource efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If optimization processing is carried out on each occasion to handle load fluctuations, then performance stability is improved, but calculation processing amount increases enormously
Solution Approach 1:
The patent applies preliminary action by predicting future load values before they occur. The prediction unit forecasts load values for subsequent time points based on historical load data, allowing the system to pre-determine resource allocation needs. This eliminates the need for reactive optimization processing on each occasion, significantly reducing calculation overhead while maintaining performance stability.
Solution Approach 2:
The patent implements periodic action by dividing time into discrete time series data points and performing prediction and allocation at regular intervals. The system collects load information at periodic time points, predicts future loads, and adjusts resource allocations periodically rather than continuously, reducing computational burden while maintaining effective load management.
2Reliability
If physical computing resource is exclusively used without sharing, then performance stability is ensured, but resource utilization efficiency decreases
Solution Approach 1:
The patent applies dynamics by making resource allocation flexible and adaptive rather than static. The system dynamically adjusts the allocation amount of physical computing resources based on predicted load values, allowing resources to be exclusively allocated when needed for performance stability while being shared during low-demand periods to improve utilization efficiency.
Solution Approach 2:
The patent implements parameter changes by varying the allocation amount of computing resources based on predicted load parameters. The system changes resource allocation parameters (allocation amount and time slot) according to predicted future loads, enabling transition between exclusive and shared usage modes to balance performance stability and resource utilization efficiency.
3Productivity
If resource allocation is adjusted frequently to match load changes, then resource utilization efficiency is improved, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary prediction unit that mediates between raw load data and resource allocation decisions. Instead of directly adjusting resources based on raw load fluctuations, the system uses the prediction unit to process and smooth load information, reducing the frequency and complexity of allocation adjustments while maintaining resource utilization efficiency.
Data Source
AI summary
A storage module holds load information obtained by classifying a load on each of virtual machines into groups depending on a fluctuation characteristic. An allocation schedule creation module predicts the load on each of the virtual machines during a predetermined period based on the load information and a rule defined for each of the groups, determining, for each of the groups, an allocation amount of the physical computing resource to be allocated to at least one of the virtual machines classified into the each of the groups and a time slot for allocation based on the predicted load on each of the virtual machines, and creating a schedule including information on the allocation amount and the time slot that have been determined. A schedule execution module allocates the physical computing resource to each of the virtual machines based on the schedule.