Cellular Device Throughput Prediction via Classification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods fail to accurately predict device throughput in cellular networks, affecting resource allocation and user experience due to variations in network types and application demands.

Innovation Solution

A system that determines per-user resource share (PULSAR) by classifying devices based on resource demand and channel quality, using a prediction engine to calculate sub-frame shares and predicted throughput, allowing for better resource allocation and network management.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing throughput prediction methods are used, then resource allocation can be performed, but the prediction accuracy is insufficient due to variations in network types and application demands

Engineering Contradiction:

Improvethroughput prediction accuracyVSAvoidadaptability to different network types and applications

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent changes key parameters by classifying devices into different types (video streaming, web browsing, file download, VoIP) based on their resource demand patterns and application characteristics. Each device type is assigned different weights for resource share calculation, enabling accurate throughput prediction adapted to specific network conditions and application requirements

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent segments the network user base into distinct device categories based on application behavior and resource consumption patterns. By dividing devices into specific types (video, web, file, voice) and applying type-specific classification, the system achieves both prediction accuracy and adaptability to different network scenarios

Inventive Principle:

Principle #1Segmentation

2Productivity

If per-user resource share calculation is implemented, then resource allocation improves, but system complexity increases due to device classification and prediction requirements

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidsystem complexity for classification and prediction

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent applies partial action by implementing classification only for resource demand parameters rather than full device characterization. The system calculates resource share based on weighted parameters specific to each device type, achieving improved resource allocation without requiring complete device profiling, thus balancing productivity gain with acceptable system complexity

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent performs preliminary device classification and resource share calculation before actual data transmission. By pre-determining device types and computing resource shares in advance based on classification, the system streamlines real-time resource allocation and reduces operational complexity during active communication

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10652756B2Device throughput determination

Publication Date: 2020.05.12 HEWLETT PACKARD ENTERPRISE DEV LP

AI summary

In one example, a system for device throughput determination includes a monitor engine to monitor an average packet size at a node of a cellular network designated for a device and a channel quality between the device and the node of the cellular network, a classification engine to classify the device based on the monitored average packet size at the node designated for the device and the channel quality, a prediction engine to predict a throughput for the device over a quantity of future transmission time intervals (TTI) based on the classification of the device and a number of resources requested by the device over a period of time.