NWDAF OAM Machine Learning Model Specification for 5G Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 5G communication systems face challenges in maintaining service quality due to the limitations of vendor-provided machine learning models for analytics information, which may not be tailored to the specific characteristics and usage patterns of the communication network, leading to suboptimal prediction accuracy and network operation.
Innovation Solution
The integration of a Network Data Analytics Function (NWDAF) and an Operation, Administration, and Maintenance (OAM) system, where the OAM system specifies and manages custom machine learning models for the NWDAF to generate analytics information, allowing for the use of models tailored to the network's specific requirements and characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vendor-provided machine learning models are used for analytics information generation, then device complexity is reduced and ease of operation is improved, but prediction accuracy and service quality deteriorate due to lack of customization to network-specific characteristics
Solution Approach 1:
The system is segmented into distinct functional components: the OAM system handles model specification and customization, while the NWDAF handles analytics generation. This segmentation allows the complexity of custom model management to be isolated in the OAM system, while the NWDAF maintains operational simplicity. The analytics function is further segmented into model-specific modules that can be independently configured and updated without affecting the entire system.
Solution Approach 2:
The OAM system performs preliminary actions by specifying and configuring custom machine learning models before the analytics generation process begins. This preliminary configuration step allows the system to prepare network-specific models in advance, ensuring high prediction accuracy when the NWDAF generates analytics information, without requiring complex real-time model adjustments during operation.
2Reliability
If custom machine learning models tailored to network characteristics are implemented, then analytics information accuracy is improved, but system complexity and configuration difficulty increase
Solution Approach 1:
The OAM system acts as an intermediary between the network operator's requirements and the NWDAF's analytics generation process. It handles the complex task of specifying and managing custom machine learning models, translating operator needs into configured model parameters. This intermediary role shields the NWDAF from configuration complexity while ensuring service quality through customized models.
Solution Approach 2:
The system enables self-service through automated model specification and configuration processes in the OAM system. Once custom models are defined, the system automatically manages their deployment and updates to the NWDAF, reducing manual configuration efforts and making the system easier to operate while maintaining high reliability through tailored models.
3Adaptability or versatility
If vendor-provided models are used, then ease of manufacture and deployment is improved, but adaptability to specific network characteristics and usage patterns deteriorates
Solution Approach 1:
The system implements dynamic adaptability through the OAM system's ability to specify and update custom machine learning models based on changing network characteristics and usage patterns. The NWDAF can dynamically adjust its analytics generation process to use the most appropriate custom models for current network conditions, while the overall system maintains ease of deployment through standardized interfaces and automated configuration processes.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to various embodiments, a 5G communication system is described comprising a Network Data Analytics Function and an Operation, Administration and Maintenance system, wherein the Operation, Administration and Maintenance system is configured to specify a machine learning model for the Network Data Analytics Function to use to generate analytics information and wherein the Network Data Analytics Function is configured to receive a specification of the machine learning model and use the machine learning model to generate analytics information.