Dynamic Network Resource Allocation via Neural Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network resource allocation methods for network slices are static, leading to waste of resources and reduced system capacity due to excessive allocation, violating Service Level Agreements (SLA) and resulting in inefficient resource utilization.
Innovation Solution
A dynamic network resource allocation method based on network slicing, which involves training a neural network with historical resource demand data to predict resource needs and pre-allocate resources, using formulas to determine optimal node and link resource quantities, and implementing an access control algorithm to manage resource allocation and access requests dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If static resource allocation is used for network slices, then network slices can occupy guaranteed resources, but network resources are wasted due to excessive allocation
Solution Approach 1:
The patent transforms static resource allocation into dynamic allocation by using neural networks to predict resource demands and adjust allocations in real-time. The system continuously monitors actual resource usage and modifies allocation strategies dynamically, allowing network slices to receive appropriate resources based on actual needs rather than fixed predetermined amounts.
Solution Approach 2:
The patent implements feedback mechanisms where the system monitors actual resource consumption by network slices and uses this information to adjust future allocations. The neural network models are trained on historical resource usage data and continuously refined based on prediction accuracy, creating a closed-loop system that learns from past performance and improves resource allocation decisions.
2Device complexity
If static resource allocation is used for network slices, then resource allocation is simple, but system capacity is reduced due to excessive resource occupation
Solution Approach 1:
The patent segments the resource allocation problem into multiple components: different neural network models for different network slice types (e.g., eMBB, URLLC, mMTC), separate prediction models for different resource types (computing, storage, networking), and hierarchical allocation levels. This segmentation allows the complex system to be managed through modular, independent components that can be optimized separately.
Solution Approach 2:
The patent changes allocation parameters dynamically based on predicted resource demands, slice priorities, and system conditions. Instead of fixed allocation parameters, the system adjusts computing resources, storage resources, and networking resources as variables that respond to changing conditions, enabling flexible optimization of system capacity while managing complexity through parameterized control.
3Speed
If static resource allocation is used for network slices, then allocation process is fast, but resource utilization efficiency is low
Solution Approach 1:
The patent performs preliminary resource allocation based on neural network predictions before actual resource demands materialize. By predicting future resource needs and pre-allocating resources accordingly, the system prepares allocation decisions in advance, maintaining fast response times while improving utilization efficiency by avoiding both over-allocation and under-allocation that plague static systems.
Data Source
AI summary
A dynamic network resource allocation method based on network slicing is provided. A historical resource demand dataset of an accessed network slice is inputted into a first neural network for training. Based on a trained first neural network and the historical resource demand of the accessed network slice, a resource demand prediction information corresponding to the accessed network slice in a first prediction time period is determined. Resources are pre-allocated to the accessed network slice based on the resource demand prediction information, and resources are allocated to the accessed network slice when the first prediction time period arrives. In this way, the service provider can reasonably allocate network resources for network slices without violating the SLA, thus avoiding the waste of network resources.