Inferential KPI Estimation for CSFB Network Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network monitoring techniques require complex and expensive data correlation across various network interfaces to determine Key Performance Indicators (KPIs), which is resource-intensive and inefficient, especially in scenarios like Circuit Switched FallBack (CSFB) in LTE networks.
Innovation Solution
Employing inferential statistical techniques to estimate KPIs by monitoring specific messages on network interfaces, such as S1-MME and SGs interfaces, and defining inferred ratios to calculate successful call counts and success rates without the need for end-to-end data correlation, thereby reducing monitoring complexity and resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional data correlation techniques are used to determine KPIs across network interfaces, then accurate KPI measurements are achieved, but system complexity and resource consumption increase significantly
Solution Approach 1:
The patent extracts only the essential monitoring points from the complex network interface data. Instead of correlating all data across multiple interfaces, it selectively monitors specific messages (Extended Service Request, Initial Context Setup Request, UE Context Modification Request) at strategic locations, removing unnecessary correlation complexity while maintaining measurement accuracy.
Solution Approach 2:
The patent introduces an intermediary statistical model that bridges the gap between limited monitored data and comprehensive KPI measurement. By using statistical inference as an intermediary, the system can estimate overall network performance metrics from a subset of monitored interfaces, avoiding the need for complex end-to-end data correlation.
2Loss of information
If comprehensive data correlation is performed across all network interfaces, then complete network scenario coverage is achieved, but bandwidth and processing resources are excessively consumed
Solution Approach 1:
The patent segments the network monitoring task into independent, manageable components. Each network interface (S1-MME, SGs) is monitored separately for specific message types, and statistical models are applied independently to each segment. This segmentation allows the system to cover comprehensive network scenarios while consuming minimal processing resources, as each segment can be analyzed independently without full correlation.
Solution Approach 2:
The patent applies partial monitoring action by selectively observing only the critical messages that indicate call setup and handover events, rather than correlating all data traffic. This partial action approach captures sufficient information to determine KPIs for complex scenarios like CSFB and handovers without the excessive resource consumption of comprehensive data correlation.
3Measurement precision
If traditional correlation methods are used to account for information from multiple network interfaces, then accurate service scenario analysis is achieved, but implementation cost and complexity increase
Solution Approach 1:
The patent enables the monitoring system to self-determine KPIs through statistical inference without requiring complex external correlation infrastructure. The system autonomously collects messages from monitored interfaces, applies statistical models locally, and generates KPI measurements independently, greatly simplifying implementation while maintaining accurate service scenario analysis for cases like CSFB and handovers.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Network monitoring devices and network monitoring techniques are disclosed and use inferential statistical approaches to determine one or more Key Performance Indicators, particularly for Circuit Switched FallBack (CSFB) scenarios. For example, the network monitoring node monitors a plurality of ciphered or clear text messages for a network interface in a communication network, determines a Mobile Terminating (MT) count for MT calls and a Mobile Originating (MO) count for User Equipment (UE) from at least one message of the plurality of messages and defines an inferred MT ratio based on the MT count and the MO count for a total number of messages. The network monitoring device further applies the inferred MT ratio to a CSFB idle mode count for UE and/or a CSFB active mode count for UE to yield an estimated total number of CSFB idle mode successes and an estimated total number of CSFB active mode successes, respectively.