Centralized Policy Engine for O-RAN Policy Conflict Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing O-RAN architecture lacks flexibility in managing and evaluating policies created by third-party rApps, leading to potential network disturbances and outages due to the absence of policy evaluation for complex use cases before implementation.
Innovation Solution
A centralized policy engine (CPE) within the SMO framework centralizes policy management, performing policy evaluation and correlation before action, integrating both user-defined and third-party rApp policies, and managing imperative policy implementation across RAN, transport, and core network functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If policies are implemented directly by third-party rApps without centralized evaluation, then policy implementation speed is improved, but network reliability deteriorates due to potential network disturbances and outages
Solution Approach 1:
The centralized policy engine performs policy evaluation and validation before policies are implemented in the network. This preliminary action assesses policy correctness, checks for conflicts, and validates policy parameters before deployment, preventing network disturbances while maintaining implementation efficiency through automated pre-checks
Solution Approach 2:
A centralized policy engine is introduced as an intermediary component between policy creation and policy implementation. This mediator evaluates, validates, and manages policies before they reach the network, ensuring reliability while maintaining a structured workflow that preserves implementation speed through efficient automated assessment
2Reliability
If centralized policy management is implemented, then network reliability is improved through policy evaluation, but device complexity increases due to additional centralized management infrastructure
Solution Approach 1:
The centralized policy engine performs multiple functions including policy evaluation, validation, conflict detection, and management within a single integrated platform. This multi-functional approach consolidates what could be multiple separate components into one system, reducing overall complexity while maintaining comprehensive policy management capabilities
Solution Approach 2:
Policy evaluation, validation, and management functions are merged into a single centralized policy engine that operates within the existing SMO framework. This consolidation integrates policy management into the existing orchestration infrastructure rather than creating separate distributed management systems, reducing system complexity while improving reliability
3Adaptability or versatility
If policies from multiple sources are integrated in a centralized engine, then adaptability is improved, but difficulty of detecting and measuring increases due to complex policy correlation requirements
Solution Approach 1:
The centralized policy engine implements feedback mechanisms that automatically track, monitor, and report on policy performance and interactions. This feedback system provides visibility into policy correlations and their effects, making complex multi-source policy management measurable and manageable while maintaining high adaptability
Data Source
AI summary
A method, performed by a centralized policy engine within a Service Management and Orchestration (SMO) framework of a telecommunications network, of managing policies for optimizing the telecommunications network is provided. The method includes: receiving, by the centralized policy engine, at least one policy, the at least one policy including a policy created by an application hosted in a radio access network (RAN) Intelligent Controller (RIC); receiving, by the centralized policy engine, data of one or more network functions, the data including at least one of performance data and event data of the one or more network functions; evaluating the at least one policy based on the received data; and controlling to implement one or more of the at least one policy in at least a RAN, based on the evaluating.