Packet Service Classification for Experience-Centric 5G Scheduling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Efficient and accurate identification of application categories generating network traffic is essential for experience-centric scheduling in multimedia content streaming services, particularly in 5G communication systems, to optimize service quality.

Innovation Solution

A machine learning-based approach using classifiers at base stations to identify application categories and video player states by analyzing IP headers of incoming packets, enabling passive classification without requiring deep packet inspection or client feedback, even with encrypted packets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If deep packet inspection is used to identify application categories, then classification accuracy is improved, but computational overhead and processing time increase significantly

Engineering Contradiction:
Improveclassification accuracyVSAvoidcomputational overhead
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential features needed for classification from packet data, specifically using IP header information and selected payload features rather than performing complete deep packet inspection. This extraction approach maintains classification accuracy while significantly reducing computational overhead by processing only the most discriminative features.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The classification process is segmented into multiple stages: first identifying application category from IP headers, then determining video player state, and finally classifying video resolution. This segmentation allows the system to process information in manageable steps, reducing overall computational complexity while maintaining high accuracy through progressive refinement.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If passive classification without client feedback is implemented, then system complexity is reduced, but measurement precision of player state may deteriorate

Engineering Contradiction:
Improvesystem complexityVSAvoidplayer state detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements an implicit feedback mechanism where the base station analyzes packet timing patterns, inter-arrival times, and traffic flow characteristics to infer video player state. This passive feedback from network-side observations enables accurate player state detection without requiring additional client device complexity or active feedback channels.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent replaces direct mechanical/client-side state reporting with network-side inference based on traffic pattern analysis. By substituting the mechanical feedback channel with analytical inference from packet characteristics, the system reduces complexity while maintaining measurement precision through sophisticated pattern recognition.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If encrypted packets are processed, then privacy is protected, but classification accuracy may worsen due to limited accessible information

Engineering Contradiction:
Improveprivacy protectionVSAvoidclassification accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent extracts classification-relevant features from encrypted packet streams without requiring decryption, specifically utilizing IP header fields, packet timing information, and flow characteristics. This extraction of metadata and temporal features maintains privacy protection while achieving accurate application and video state classification through non-content-based features.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of trying to access encrypted content directly, the patent inverts the approach by classifying based on external characteristics of the encrypted stream - such as packet timing, size patterns, and protocol headers - thereby maintaining privacy while achieving classification accuracy through alternative information channels.

Inventive Principle:
Principle #13The other way round (Inversion)

4Reliability

If experience-centric scheduling is implemented, then service quality is improved, but device complexity and processing requirements increase

Engineering Contradiction:
Improveservice qualityVSAvoidprocessing requirements
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary classification of application category, video player state, and video resolution before scheduling decisions are made. By pre-processing and categorizing traffic in advance using lightweight feature extraction from IP headers, the system enables experience-centric scheduling without adding significant processing complexity during the actual scheduling execution phase.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates a universal classification framework that handles multiple service types (video streaming, general data traffic) and multiple classification dimensions (application category, player state, resolution) through a single integrated system. This multi-functional approach improves service quality across diverse applications while avoiding the need for separate specialized processing systems.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3777058B1Machine learning based packet service classification methods for experience-centric cellular scheduling
Publication Date: 2025.11.26 SAMSUNG ELECTRONICS CO LTD
  • EP3777058B1 patent drawingFigure 1~2
  • EP3777058B1 patent drawingFigure 3
  • EP3777058B1 patent drawingFigure 4

AI summary

The present disclosure relates to a communication method and system for converging a 5th-Generation (5G) communication system for supporting higher data rates beyond a 4th-Generation (4G) system with a technology for Internet of Things (IoT). The present disclosure may be applied to intelligent services based on the 5G communication technology and the IoT-related technology, such as smart home, smart building, smart city, smart car, connected car, health care, digital education, smart retail, security and safety services. And methods and apparatus for classifying a traffic packet service of a traffic flow to a client device, a state and/or video resolution of a streaming video player in a client device. The apparatus includes a memory, and a communication interface configured to obtain TCP/IP or UDP /IP headers of packets to a client device, at least one processor operably connected to the memory, the at least one processor configured to detect a start of a traffic flow, and buffer a first number of packets for the traffic flow, extract features from UDP/IP or TCP/IP headers of a second number of packets belonging to the detected traffic flow, and obtain a flow service classifier for classifying a service type of the traffic flow, and classify the service type of the traffic flow by applying the extracted features to the flow service classifier.