Edge Computing Predictive Provisioning via ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing environments face challenges in predicting and optimizing resource provisioning for edge computing environments, particularly in managing infrastructure and virtualization parameters to meet service level agreements (SLAs) due to the variability and complexity of edge computing setups.
Innovation Solution
A method and system that utilize machine learning to train predictive models based on utilization parameter values from edge computing environments, allowing for the prediction of future performance and proactive reprovisioning decisions, including adjustments to infrastructure and virtualization configurations, driven by updated SLA parameters and similarity analysis with other environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional reactive provisioning methods are used in edge computing environments, then infrastructure and virtualization parameters can be adjusted based on current utilization, but service level agreements cannot be proactively met due to the variability and complexity of edge computing setups
Solution Approach 1:
The system performs preliminary actions by training machine learning predictive models using historical utilization parameter values from multiple edge computing environments before provisioning decisions are needed. These models predict future performance metrics, enabling the system to proactively adjust infrastructure and virtualization parameters in advance to meet service level agreements, rather than reacting to current conditions alone.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring utilization parameter values from edge computing environments and using these measurements to retrain and refine predictive models. The models incorporate feedback from multiple environments through similarity analysis, allowing the system to learn from diverse configurations and improve its provisioning decisions while adapting to the complexity of different edge computing setups.
2Productivity
If predictive models are trained using historical utilization data from multiple edge computing environments, then future performance can be predicted and proactive reprovisioning decisions can be made, but computational resources and time are required for model training and processing
Solution Approach 1:
The system performs model training in advance using historical utilization parameter values collected from multiple edge computing environments. By preparing predictive models beforehand, the system avoids the need for time-consuming training operations when provisioning decisions are urgently needed, thus reducing the time loss during critical decision-making moments while maintaining high provisioning efficiency.
Solution Approach 2:
The system applies partial training actions by updating models with new data at selective intervals rather than continuously retraining from scratch. This approach uses incremental learning where only necessary model adjustments are made based on new utilization data, reducing the computational time and resources required compared to complete retraining, while still improving prediction accuracy for proactive provisioning.
3Measurement precision
If utilization parameter values are obtained and analyzed from multiple edge computing environments to train predictive models, then more accurate predictions can be made, but data collection and processing complexity increases
Solution Approach 1:
The system implements a universal data collection framework that standardizes the extraction of utilization parameter values across diverse edge computing environments. This multi-functional approach allows the same data collection and processing mechanisms to work across different environment types and configurations, reducing the apparent complexity by providing a unified interface for gathering training data from multiple sources with varying characteristics.
Solution Approach 2:
The system introduces intermediary components that act as mediators between diverse edge computing environments and the predictive model training process. These intermediaries standardize and normalize utilization parameter data from different environments, handling the complexity of data collection and preprocessing centrally, thereby simplifying the overall process while maintaining high measurement precision through consistent data quality across all environments.
Data Source
AI summary
Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: iteratively obtaining utilization parameter values from first to Nth edge computing environments, training one or more predictive model by machine learning using parameter values of the utilization parameter values obtained by the iteratively obtaining, wherein the training includes training a first computing environment predictive model with use of parameter values of the utilization parameters obtained from the first computing environment by the iteratively obtaining.