Dynamic Application Instance Selection Using Network Performance Analytics

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing communication systems struggle to select the optimal server application instance for UEs based on their location and performance, leading to suboptimal traffic distribution, resilience, and communication performance.

Innovation Solution

Implementing a system that utilizes UE location and performance data analytics through the NWDAF and APMF to determine the best server application instance by considering historical performance data and network analytics, using the SDMF to make informed selection decisions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a UE is located in an edge data network service area, it receives the address of an appropriate edge-instance Server Application, but when the UE roams outside the edge data network service area, it receives the address of a default cloud-based instance, leading to suboptimal performance for roaming UEs

Engineering Contradiction:
Improvecommunication performanceVSAvoidserver selection adaptability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The server selection mechanism transitions from a static location-based approach to a dynamic analytics-driven approach. The NWDAF continuously collects performance data from multiple sources (APMF, AMF, SMF, UPF) and updates analytics in real-time, allowing the system to adapt server recommendations based on current network conditions, UE mobility patterns, and historical performance, rather than relying solely on predetermined edge service area boundaries

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements a closed-loop feedback mechanism where performance data is continuously collected from various network functions (APMF for application performance, AMF for mobility management, SMF for session management, UPF for user plane performance), analyzed by the NWDAF, and used to update server selection recommendations. This feedback loop enables the system to learn from past performance and continuously improve server instance selection for both edge and cloud deployments

Inventive Principle:
Principle #23Feedback

2Productivity

If the system uses basic location-based routing for server selection, then the implementation is simple, but the traffic distribution and resilience are suboptimal

Engineering Contradiction:
Improvetraffic distribution efficiencyVSAvoidanalytics function complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The analytics function is segmented into distinct modular components: APMF for application-level performance monitoring, AMF for mobility and location management, SMF for session and bearer management, UPF for user plane performance measurement, and NWDAF for data collection and analytics processing. Each component has a specific responsibility and interfaces through standardized protocols, allowing independent deployment, maintenance, and optimization of each segment while achieving sophisticated overall functionality

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The NWDAF acts as an intermediary between the performance data sources (APMF, AMF, SMF, UPF) and the server selection logic. It aggregates data from multiple sources, processes the information through analytics algorithms, and provides refined recommendations to the SDMF or other selection entities. This intermediary layer simplifies the overall architecture by centralizing the complex analytics processing while maintaining clean interfaces with constituent components

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If the system selects server instances without considering performance analytics, then the selection process is fast, but the average communication performance across multiple sessions is suboptimal

Engineering Contradiction:
Improveaverage communication performanceVSAvoidserver selection time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by continuously collecting and analyzing performance data in the background before actual server selection is needed. The NWDAF maintains updated analytics about server instance performance, UE mobility patterns, and network conditions, so that when a server selection decision is required, the pre-computed analytics can be quickly retrieved and applied, reducing real-time decision latency while ensuring performance-optimized selections

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4150967B1Selecting an application instance
Publication Date: 2025.08.13 LENOVO (SINGAPORE) PTE LTD
  • EP4150967B1 patent drawingFigure 1
  • EP4150967B1 patent drawingFigure 2
  • EP4150967B1 patent drawingFigure 3

AI summary

Apparatuses, methods, and systems are disclosed for selecting a server application instance. One apparatus (800) includes a network interface (840) that communicates with a plurality of network functions in a mobile communication network and a processor (805) that receives (905) a request from a first network function to provide performance analytics for a first application. Here, the request includes a present location and a requested time. Additionally, the first application comprises a group of application instances. The processor (805) generates (910) performance analytics for the first application by using a first collection of performance data and reports (915) the performance analytics to the first network function. Here, the performance analytics indicate a best application instance in the group of application instances for the present location and the requested time.