5G Network Management via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The 5G network architecture faces challenges in optimal resource utilization, network management flexibility, and ensuring service level agreements (SLAs) are met before and after network slice creation, particularly when resources are split across multiple slices.
Innovation Solution
A device utilizing machine learning for closed-loop network management processes analytics data from the core, edge, and radio access network domains to determine and perform actions that optimize resource allocation and slice management, ensuring SLAs are satisfied through hierarchical architecture and domain-specific models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If network resources are split across multiple network slices to support diverse use cases, then network versatility and service customization are improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements dynamic resource allocation where network slice resources are not statically assigned but continuously adjusted based on real-time demand. The system monitors resource usage metrics and automatically reallocates resources between slices, enabling the network to adapt to changing traffic patterns and service requirements while maintaining optimal utilization across all slices.
Solution Approach 2:
The system changes key parameters such as resource allocation ratios, bandwidth assignments, and capacity distribution dynamically based on network conditions. By adjusting these parameters in response to real-time analytics, the system optimizes resource utilization while maintaining the ability to support diverse network slice requirements.
2Productivity
If manual management methods are used for network slice creation and configuration, then implementation simplicity is maintained, but management efficiency and SLA guarantee capability deteriorate
Solution Approach 1:
The patent implements self-service automation where the network management system automatically performs slice creation, configuration, resource allocation, and optimization without requiring manual intervention. The system uses machine learning models to autonomously make decisions about resource distribution and slice management, significantly improving efficiency while handling the complexity internally.
Solution Approach 2:
The system incorporates continuous feedback loops that monitor network slice performance, resource utilization, and SLA compliance. This feedback is fed back into the management system to automatically adjust configurations and resource allocation, enabling efficient automated management while maintaining adaptability to changing conditions.
3Adaptability or versatility
If static resource allocation is used for network slices, then configuration simplicity is maintained, but flexibility for resizing and rebalancing deteriorates
Solution Approach 1:
The patent transforms static resource allocation into a dynamic system where resource assignments are continuously adjusted based on real-time network conditions and slice performance. The system automatically performs resizing and rebalancing operations, providing the flexibility needed for diverse service requirements while managing the complexity of dynamic adjustments through automated control mechanisms.
Data Source
AI summary
A device receives analytics data associated with management of a network associated with the device, core data associated with a core domain of the network, edge data associated with an edge domain of the network, and radio access network (RAN) data for a RAN associated with the network. The device processes the analytics data, the core data, the edge data, and the RAN data, with a machine learning model, to determine actions to be performed with respect to the core domain of the network, the edge domain of the network, and/or the RAN. The device causes the actions to be performed by one or more core devices associated with the core domain of the network, one or more edge devices associated with the edge domain of the network, and/or one or more RAN devices associated with the RAN.