Network Slice Load Monitoring With Cost-Based Instantiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current industry standards fail to define how to accurately monitor network slices and manage load on slices, leading to inefficient resource consumption and traffic loss due to inaccurate monitoring and management, resulting in the instantiation of additional slices based on inaccurate results.
Innovation Solution
A monitoring system that evaluates network slice performance by packet loss percentage, jitter, and packet throughput, and then assigns a score based on these metrics to determine if additional slices or slice instances are needed, or if existing slices should be switched or rejected, using a cost model and machine learning to optimize resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network slice monitoring is performed using current industry standards, then slice performance can be tracked, but the monitoring accuracy is insufficient leading to inaccurate results
Solution Approach 1:
The patent implements a feedback mechanism where the cost model continuously receives network data (packet loss, jitter, latency, throughput) and updates slice cost scores. This feedback loop enables dynamic adjustment of resource allocation decisions based on real-time performance metrics, improving both monitoring accuracy and reliability of load management
Solution Approach 2:
The patent replaces traditional mechanical threshold-based monitoring with an AI-based cost model that uses machine learning algorithms. This substitution transforms rigid rule-based systems into adaptive intelligent systems that can accurately assess slice performance and predict future states, significantly improving measurement precision and decision reliability
2Productivity
If additional slices are instantiated based on inaccurate monitoring results, then network capacity increases, but resource consumption becomes inefficient
Solution Approach 1:
The cost model performs preliminary evaluation of slice performance using network data before making instantiation decisions. By predicting future slice states and calculating cost scores in advance, the system avoids unnecessary slice creation, ensuring that additional slices are only instantiated when truly needed, thus maintaining productivity while preventing resource waste
Solution Approach 2:
The patent dynamically changes monitoring parameters and cost weights based on network conditions and service requirements. This allows the system to adapt resource allocation strategies to different scenarios, optimizing the balance between network capacity expansion and resource consumption efficiency
3Device complexity
If traditional load management methods are used, then implementation is simple, but traffic loss occurs due to inaccurate load assessment
Solution Approach 1:
The patent introduces an intermediary cost model layer between network monitoring and load management decisions. This cost model acts as a mediator that processes network data, calculates cost scores, and provides actionable insights, thereby reducing traffic loss through more accurate load assessment while maintaining manageable system complexity through modular architecture
Data Source
AI summary
A device may receive network data identifying uplink/downlink packet loss percentage, uplink/downlink jitter, uplink/downlink latency, and uplink/downlink packet throughput associated with a slice of a network, and may set an uplink value for each of the uplink packet loss percentage, the uplink jitter, the uplink latency, and the uplink packet throughput. The device may multiply the uplink values by corresponding uplink weights to calculate an uplink cost, and may set a downlink value for each of the downlink packet loss percentage, the downlink jitter, the downlink latency, and the downlink packet throughput. The device may multiply the downlink values by corresponding downlink weights to calculate a downlink cost, and may calculate a total cost based on the uplink cost and the downlink cost. The device may cause another slice, with the same attributes as the slice, to be instantiated when the total cost satisfies a threshold for a time period.