Network Slice Assignment via Multi-Dimensional Attribute Matrix Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In advanced cellular networks like 5G New Radio, efficiently assigning user equipment to network slices that meet varying service level agreement (SLA) requirements becomes increasingly difficult as the number of services and slices grows, leading to uneven loading and suboptimal performance.
Innovation Solution
A method involving a network slice management system that creates a multi-dimensional attribute matrix, uses machine learning algorithms for clustering network slices, performs classification and threshold-based analysis to select the most suitable slice for client requests, and provides service using the selected slice, optimizing resource allocation and cost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of network slices increases to meet diverse SLA requirements, then service versatility improves, but slice assignment complexity increases
Solution Approach 1:
The patent segments the large set of network slices into multiple classes based on service characteristics and SLA requirements. By categorizing slices into distinct classes (e.g., enhanced mobile broadband, ultra-reliable low-latency communication, massive machine type communication), the system simplifies the assignment process while maintaining the ability to serve diverse services. This segmentation reduces the complexity of matching UEs to appropriate slices by limiting the search space to relevant slice classes.
Solution Approach 2:
The patent changes the parameter space by representing network slices through key attributes and parameters (e.g., bandwidth, latency, reliability requirements) rather than treating each slice as a unique entity. This parameter-based representation enables efficient comparison and matching between UE requirements and slice characteristics, reducing assignment complexity while preserving service versatility.
2Device complexity
If manual slice assignment methods are used, then system simplicity is maintained, but assignment efficiency deteriorates
Solution Approach 1:
The patent implements self-service through automated slice assignment mechanisms where the network management system automatically classifies UEs and assigns them to appropriate network slices based on their service requirements. This eliminates the need for manual intervention while maintaining system simplicity through rule-based or machine learning-driven automated decision-making processes.
Solution Approach 2:
The patent incorporates feedback mechanisms where the network management system continuously monitors UE service requirements, slice utilization, and performance metrics. This feedback enables dynamic adjustment of slice assignments, improving efficiency by adapting to changing conditions while maintaining manageable system complexity through automated control loops.
3Device complexity
If traditional assignment methods are used, then implementation simplicity is maintained, but resource allocation optimality deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-classifying network slices into distinct categories based on their service characteristics and SLA requirements before actual UE assignment occurs. This pre-organization of slices into classes enables more optimal resource allocation when UEs are assigned, as the matching process can efficiently identify suitable slice classes based on UE requirements, improving allocation optimality without significantly increasing implementation complexity.
Data Source
AI summary
Various network slice assignment arrangements are presented. A network slice management system can create a multi-dimensional attribute matrix that defines network slices that operate on a physical cellular network. Network slices that operate on the physical cellular network can be clustered based on attributes to create slice classes. A classification process to determine a slice class that most closely matches a cellular network service request may be performed. A network slice can then be selected from the slice class and be used for providing cellular network service.