Namespace Resource Forecasting With Graph Neural Dependencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning forecasting models for namespace resource consumption in containerized applications fail to accurately predict usage due to hidden dependencies among pods, namespaces, and clusters, leading to potential resource overconsumption and inefficiencies.
Innovation Solution
Implementing multivariate time series forecasting with a graph neural network to learn hidden spatial and temporal relationships between variables at the namespace level, using normalized timeseries datasets to generate accurate resource consumption predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing machine learning forecasting models are used for namespace resource consumption, then resource consumption can be forecasted, but the predictions are inaccurate due to hidden dependencies among pods, namespaces, and clusters
Solution Approach 1:
The patent introduces a graph neural network as an intermediary component that mediates between the raw timeseries data and the forecasting output. This intermediary learns and captures the hidden dependencies among pods, namespaces, and clusters through graph structure representation, enabling accurate predictions while maintaining reliability by accounting for complex inter relationships that conventional models miss
2Productivity
If conventional forecasting models are used, then the system remains simple, but resource overconsumption and inefficiencies occur due to inaccurate predictions
Solution Approach 1:
The patent transforms the forecasting approach by changing key parameters: switching from conventional statistical models to graph neural networks, changing the data representation from simple timeseries to graph-structured data, and changing the prediction output from single-value forecasts to probability distributions. These parameter changes enable the system to handle complex dependencies while improving resource allocation efficiency through more accurate predictions
Data Source
AI summary
The technology described herein is directed towards determining predicted and/or actual namespace resource consumption in an automated system for deployment, scaling, and management of containerized applications, such as in a Kubernetes® system. Given time series data representative of cluster-level resource consumption history at a percentage scale, resource consumption history for every namespace in the cluster at an absolute scale, and resource consumption history for each individual pod in the cluster at an absolute scale, multivariate time series forecasting with graph neural networks learns the hidden (dynamic and time variant) variable dependencies during a forecasting process that includes graph convolution followed by temporal convolution. The result is a forecast of a namespace's resource consumption.