Custom Network Analytics Model Training for Scalable Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network analytics systems are inflexible and not scalable, as they rely on a predefined list of analytics services that may not meet the diverse needs of service consumers, particularly in 5G and beyond networks.
Innovation Solution
The proposed solution involves an apparatus and method that enable custom analytics services by allowing service consumers to request specific analytics, with the system determining and selecting a suitable model training entity to create or obtain a trained model capable of providing the requested analytics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a predefined list of analytics services is used, then system simplicity is maintained, but adaptability to diverse service consumer needs deteriorates
Solution Approach 1:
The system dynamically adapts between two operational modes: using a predefined analytics service list for standard requests (maintaining simplicity) and enabling custom analytics model training when service consumer needs are not met (providing adaptability). This dynamic switching resolves the contradiction by adjusting system behavior based on request type.
Solution Approach 2:
The analytics service provision is segmented into two distinct pathways: (1) predefined analytics services from a fixed list that maintain system simplicity, and (2) custom analytics services with model training capabilities that provide adaptability. This segmentation allows each pathway to optimize for its specific purpose without compromising the other.
2Adaptability or versatility
If custom analytics services are enabled, then service versatility is improved, but system complexity increases
Solution Approach 1:
The system introduces an intermediary mechanism (the model training entity and selection module) that bridges the gap between service consumer requests and analytics provision. This intermediary handles the complexity of custom model training and selection, shielding service consumers from underlying system complexity while enabling versatile custom analytics services.
Solution Approach 2:
The system enables self-service capabilities where model training entities can autonomously train and provide custom analytics models based on service consumer requests. This self-service approach reduces the need for manual system configuration and management, thereby managing complexity while maintaining service versatility.
3Productivity
If predefined analytics services are used, then system scalability is limited, but operational simplicity is maintained
Solution Approach 1:
The system achieves multi-functionality by combining a predefined analytics service list (for operational simplicity) with a model training capability (for scalability). The same system infrastructure supports both standard predefined services and customized analytics services, allowing the system to scale to diverse requirements without sacrificing operational simplicity for either pathway.
Data Source
AI summary
An apparatus for use by a communication network element or communication network function configured to operate as an analytics entity and having an analytics logical function. The apparatus is caused to: receive a request, from a service consumer, for provision of a custom analytics service, process data retrieved from the request for determining whether a model for providing the requested custom analytics service is prepared, and in case the determination is negative, determine and select a model training entity having a model training logical function which is able to create a model for custom analytics and has access to data required for the requested custom analytics service, request, from the selected model training entity, to obtain a trained model capable of providing the requested custom analytics service and forward the information specifying the custom analytics service to the selected model training entity.