CAT-LSTM Model for VNF Resource Demand Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting demand for virtual network function (VNF) resources in network management are inaccurate due to the lack of consideration for environmental and overall network information, requiring extensive expertise and recollection of resource data upon changes, and necessitate multiple machine learning models for each VNF.
Innovation Solution
A machine learning-based method using a content and aspect embedding attentive target dependent long short-term memory (CAT-LSTM) model that predicts VNF resource demand by incorporating feature data, attention learning, and auxiliary data from a service function chain (SFC), allowing for accurate resource allocation and management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If resource information of individual VNF is used for prediction, then the prediction model is simple, but the prediction accuracy is low due to lack of environmental information
Solution Approach 1:
The patent merges resource information from multiple VNFs within the same SFC into a unified prediction model. Instead of predicting each VNF's resource demand independently, the system combines their resource information to perform joint prediction, thereby improving accuracy while avoiding the need for multiple separate models.
Solution Approach 2:
The patent creates a universal prediction model that can handle multiple VNFs simultaneously. This single model serves the function of predicting resource demand for all VNFs in an SFC, replacing what would otherwise require multiple individual prediction models, thus reducing overall system complexity.
2Productivity
If resource information of each VNF is collected separately, then the prediction can be performed individually, but extensive expertise and time are required for data recollection upon network changes
Solution Approach 1:
The system merges the data collection process for multiple VNFs into a single unified collection operation. When predicting resource demand for VNFs in an SFC, the system collects resource information for all target VNFs simultaneously in one process, rather than collecting data for each VNF separately, thereby reducing the time and effort required during network changes.
3Measurement precision
If multiple machine learning models are used for each VNF, then individual prediction can be performed, but the system complexity and operational costs increase
Solution Approach 1:
The patent merges multiple individual VNF prediction models into a single unified prediction model. This model takes resource information from multiple VNFs as input and produces predictions for all of them, thereby reducing the total number of models required while maintaining prediction accuracy through the use of combined environmental information.
Solution Approach 2:
The unified prediction model serves multiple functions by predicting resource demand for all VNFs within an SFC simultaneously. This multi-functional model replaces what would otherwise require multiple specialized models, reducing system complexity and operational overhead while maintaining comprehensive prediction coverage.
Data Source
AI summary
The present invention relates to a technique in which demand prediction of resources of virtual network functions (VNFs) that provide a core technology in a network virtualization environment is performed using machine learning technology. In the present invention, in order to predict VNF resource information, not only are the resources of the VNFs as data but also information of surrounding VNFs that are directly or indirectly related are used, and prediction is possible even in a dynamically changed network environment. In addition, service function chain (SFC) data among various pieces of network information is used to reduce a time required for machine learning according to a size of an entire network.