Separating Uplink Downlink Classifiers for Traffic Steering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current traffic steering technologies in mobile networks face high overhead and resource consumption due to passing both uplink and downlink service traffic through the same classifier, leading to inefficient processing and potential bottlenecks, without adequate analysis of service characteristics and optimization of downlink traffic processing.
Innovation Solution
A service traffic steering method and device that separates the processing of uplink and downlink traffic by using distinct classifiers, where the uplink classifier identifies and sends a steering path identifier and description information to the downlink classifier only when necessary, optimizing the routing and reducing unnecessary resource consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If both uplink and downlink service traffic pass through the same classifier, then traffic identification can be performed, but processing overhead increases and resource consumption increases
Solution Approach 1:
The patent divides the single classifier into separate uplink classifier and downlink classifier, segmenting the traffic processing function according to direction. This allows parallel processing of uplink and downlink traffic, reducing the processing burden on a single classifier while maintaining identification accuracy.
Solution Approach 2:
The patent introduces a new dimension for traffic classification by separating classifiers based on traffic direction (uplink vs. downlink). This dimensional separation enables independent optimization of each classifier's processing capacity and reduces contention for shared resources.
2Device complexity
If a single classifier processes all traffic, then implementation is simple, but a bottleneck easily occurs and reliability decreases
Solution Approach 1:
By segmenting the single classifier into multiple direction-specific classifiers, the patent eliminates the single-point failure risk. If one classifier fails or becomes overloaded, the other continues to operate, improving system reliability without significantly increasing overall structural complexity.
Solution Approach 2:
The patent changes the operational parameters of the classification system by introducing direction-specific processing rules and separate classifier instances. This parameter change allows the system to handle different traffic types more reliably while maintaining manageable complexity through standardized interfaces.
3Measurement precision
If downlink traffic is identified by PGW or TDF and then returned to Gi-LAN for processing, then traffic can be processed, but the route becomes roundabout and processing time increases
Solution Approach 1:
The downlink classifier performs traffic identification and classification in advance, before traffic enters the Gi-LAN for value-added service processing. This preliminary action eliminates the need for traffic to return to the Gi-LAN for re-processing, directly reducing processing time while maintaining identification accuracy.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A service traffic steering method, device and system are provided. The method includes that: a steering path identifier of a downlink service traffic is identified according to a service traffic steering policy; and the steering path identifier of the downlink service traffic and description information of the downlink service traffic are sent to a downlink classifier to enable the downlink classifier to mark the downlink service traffic according to the steering path identifier of the downlink service traffic, an uplink classifier and the downlink classifier are deployed in different network devices.