Service-Based AI/ML Model Lifecycle for RAN Intelligent Controllers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a lack of clear specifications on the necessary AI/ML functions, their locations, and the mechanisms for implementing, training, certifying, registering, and deploying AI/ML models in RICs within the Open RAN architecture, particularly in the context of the O-RAN Alliance's Non-RT and Near-RT RIC architectures.
Innovation Solution
A service-based AI/ML architecture is introduced, comprising AI/ML Management and Exposure functions, model training, certification, registration, deployment, and inference functions, along with an inventory, which can be implemented in various configurations including Non-RT RIC, Near-RT RIC, or a combination, and includes integrated and separated procedures for training, certification, registration, and deployment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If AI/ML functions are integrated into RIC architecture, then model training and deployment capabilities are improved, but system complexity increases
Solution Approach 1:
The patent segments the AI/ML functionality into separate service functions (training function, verification function, registration function, deployment function) that can be independently implemented and managed. This allows the complex AI/ML capabilities to be modularized within the RIC architecture, reducing overall system complexity while maintaining versatility.
Solution Approach 2:
The patent creates a universal service-based architecture where a single RIC can support multiple AI/ML functions through standardized interfaces. The management function can orchestrate different service functions (training, verification, registration, deployment) making the system multi-functional without requiring separate dedicated systems for each capability.
2Ease of operation
If service-based architecture is implemented for AI/ML functions, then ease of management is improved, but interface complexity increases
Solution Approach 1:
The patent introduces a management function as an intermediary that orchestrates communication between different service functions (training, verification, registration, deployment). This intermediary layer simplifies management by providing a centralized control point while the service functions communicate through standardized interfaces, reducing the complexity of direct interactions between multiple components.
3Productivity
If comprehensive AI/ML functions are deployed, then productivity is improved, but loss of time increases
Solution Approach 1:
The patent implements preliminary verification and registration of AI/ML models before deployment to the network. The verification function validates model correctness and the registration function records model metadata in advance, so that when models need to be deployed, the time-consuming verification and registration steps are already complete, reducing overall deployment time while maintaining productivity.
Data Source
AI summary
A system for supporting artificial intelligence/machine learning (AI/ML) model functions using a service-based architecture in a radio access network (RAN) intelligent controller (RIC) is provided. The system includes a first function for managing AI/ML functions, and for exposing management and exposure services for the AI/ML functions. The system also includes a second function for providing services for deploying the AI/ML models in the at least one RIC, and a repository for storing the AI/ML models. The first function, the second function, and the repository are connected with the service-based architecture.