Distributed Learning for 5G Network Slice Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The management of persistent network slices in 5G wireless communication systems is complex due to the scale and diversity of devices, making allocation, scheduling, and resource management difficult, especially across multiple access points and contexts.
Innovation Solution
A distributed learning system with AI/ML components and blockchain ledger is employed to facilitate the allocation, scheduling, and management of network slices, allowing for flexible resource distribution, personalized slice customization, and secure, persistent storage of slice usage data across multiple contexts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If network slices are allocated to millions or billions of devices, then service coverage and capacity are improved, but system complexity and management difficulty increase
Solution Approach 1:
The patent segments the network slice management into multiple distributed learning agents operating at different network levels (access point, local network, core network). Each agent independently manages slices within its domain using local machine learning models, dividing the overwhelming global management task into manageable local decisions. This segmentation allows billions of slices to be handled through distributed intelligence rather than centralized control.
Solution Approach 2:
The system implements self-service through autonomous learning agents that automatically allocate, optimize, and manage network slices using machine learning without human intervention. The distributed agents continuously learn from local network conditions and autonomously make slicing decisions, enabling the system to self-manage the complex allocation across billions of devices. Blockchain technology further enables self-verification and self-preservation of slice configurations.
2Productivity
If distributed learning agents are deployed across multiple access points, then local resource optimization is improved, but communication overhead and coordination complexity increase
Solution Approach 1:
The management function is segmented across multiple hierarchical levels with autonomous learning agents at each level handling local decisions independently. Access point-level agents manage local slice allocations without needing to communicate every decision to the core network, significantly reducing communication overhead while maintaining global optimization through periodic model updates via blockchain.
Solution Approach 2:
Blockchain technology serves as an intermediary layer that enables efficient coordination between distributed learning agents. Instead of direct peer-to-peer communication which would generate excessive overhead, the blockchain provides a standardized interface for agents to share learnings, verify allocations, and maintain consistency across the distributed system with minimal communication requirements.
3Reliability
If persistent network slice data is stored across distributed nodes, then system robustness and availability are improved, but data consistency and security management become more difficult
Solution Approach 1:
The blockchain ledger provides universal data management functionality across all distributed nodes, serving simultaneously as a consistent storage mechanism, security verification system, and coordination platform. This multi-functional approach simplifies data management complexity by providing a single standardized interface that handles consistency, security, and reliability requirements across all nodes without requiring separate mechanisms for each function.
Solution Approach 2:
The system implements continuous feedback loops where learning agents monitor network conditions, evaluate slice performance, and automatically adjust allocations based on real-time data. Blockchain provides verifiable feedback mechanisms that allow agents to confirm successful data replication and consistency across nodes, enabling robust distributed storage while maintaining simplicity through automated verification and adjustment.
Data Source
AI summary
The technologies described herein are generally directed to facilitating the allocation, scheduling, and management of network slice resources. According to an embodiment, a system can comprise a processor and a memory that can store executable instructions that, when executed by the processor, facilitate performance of operations. The operations can include selecting a resource configuration for a network slice based on characteristics of a user device and historical data related to the user device, resulting in a selected resource configuration. The operations can further include facilitating communicating resource configuration data representative of the selected resource configuration for the network slice to a network device for allocation to the user device connected to the network device. The operations can further include facilitating allocating resources to the network slice in accordance with the selected resource configuration.