Network Slice Admission Control Using ML Demand Load Shaping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing slice admission control mechanisms in 5G networks face challenges in optimizing network slice allocation due to unpredictable and uncoordinated user requests, leading to resource inefficiencies and collisions, with methods failing to account for flexible application requirements and resulting in high computational complexity and slow convergence times.
Innovation Solution
A slice admission control method utilizing a machine learning model to determine slice utility gains based on network resource demand loads, enabling coordinated and predictable resource allocation by shaping slice requests and optimizing resource usage through a slice controller.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing slice admission control mechanisms process uncoordinated and random user requests, then the network can handle diverse service types, but resource efficiency deteriorates due to collisions and unpredictable resource allocation
Solution Approach 1:
The patent introduces a preliminary request shaping phase where the network controller provides guidance information to user equipment before actual slice requests are submitted. This preliminary action coordinates user requests to avoid collisions and optimize resource allocation, transforming the random request pattern into a coordinated one while maintaining service diversity
Solution Approach 2:
The patent implements a feedback mechanism where the network controller monitors resource allocation status and provides guidance information back to user equipment. This feedback loop enables continuous optimization of request coordination, allowing the system to adapt to changing network conditions while maintaining efficient resource utilization
2Measurement precision
If slice admission control algorithms consider all network slices and states to ensure comprehensive resource allocation, then allocation accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the complex admission control problem into two distinct phases: request shaping (coordination) and admission decision-making. This segmentation allows each phase to focus on specific aspects of resource allocation, reducing the computational burden on any single algorithm while maintaining comprehensive consideration of network slices and states in the decision phase
3Reliability
If slice admission control processes all incoming requests with fixed constraints to ensure reliability, then service reliability improves, but convergence time increases due to extensive processing
Solution Approach 1:
The patent performs preliminary request shaping that coordinates and optimizes user requests before they reach the admission control stage. This preliminary processing reduces the number of conflicting requests that require extensive processing, thereby maintaining service reliability through comprehensive constraint checking while reducing convergence time by eliminating redundant processing of coordinated requests
Data Source
AI summary
A slice admission control method performed by a slice controller in a communication network system deploying a plurality of network slices and comprising:a) receiving a slice request containing slice request parameters from a user equipment,b) obtaining network parameters related to at least an overall state of network resources in the plurality of network slices,c) determining, for each network slice:a slice service variable related to required network resources,a slice utility gain to be expected from a utilization of the required network resources,a slice demand load related to consumed network resources in the network slice,d) transmitting the slice demand loads to the user equipment,e) feeding the slice demand loads to the machine learning model,f) based on the slice utility gain, determining a decision on the admission of the slice request.