Web Quality Estimation Using Precomputed Content Characteristics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current web quality estimation methods are inaccurate due to neglecting throughput variations in real networks and require high calculation costs and frequent operation log measurements, especially when web content changes frequently.
Innovation Solution
A web quality estimation device that estimates processing time based on input network qualities and content characteristics, using a content characteristic storage unit and web quality estimation unit to calculate web quality without simulating network qualities or acquiring operation logs, and optionally includes a content characteristic estimation unit to derive content characteristics from feature values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If web quality is estimated by network quality simulation based on operation log information measured in advance, then both latency and throughput can be taken into consideration, but high calculation cost is required
Solution Approach 1:
The patent pre-calculates and stores quality metrics for multiple network quality combinations in a quality information storage unit during system initialization. This preliminary action allows the estimation device to quickly retrieve pre-computed quality values during operation without performing expensive real-time simulations, thus resolving the contradiction between measurement precision and calculation cost.
Solution Approach 2:
The patent creates a copy of quality information from simulations performed in advance and stores it in a database structure. Instead of performing the actual simulation during estimation, the system uses these pre-generated copies of quality data, which significantly reduces calculation cost while maintaining the accuracy benefits of considering both latency and throughput.
2Measurement precision
If operation log information is measured every time web content changes, then accurate web quality estimation can be achieved, but high cost for measurement is required
Solution Approach 1:
The system pre-calculates quality information for multiple network quality combinations and stores it in advance. When web content changes, instead of measuring operation logs again, the system retrieves the appropriate pre-computed quality information based on the new content characteristics, eliminating the need for repeated measurements while maintaining accuracy.
Solution Approach 2:
The patent uses content identification information and network quality parameters to index into pre-computed quality tables. By changing the approach from measuring operation logs to retrieving quality information based on content and network parameters, the system achieves accurate estimation without repeated measurements, thus resolving the contradiction between precision and time loss.
3Device complexity
If throughput is not taken into consideration in web quality estimation, then estimation can be simplified, but accuracy of web quality estimation deteriorates in real network environments
Solution Approach 1:
The patent pre-computes quality information considering both latency and throughput for multiple network quality combinations and stores it in a quality information storage unit. This preliminary computation captures the complex relationships between network qualities and web quality, allowing the system to consider both throughput and latency without increasing real-time estimation complexity, thus resolving the contradiction between device complexity and measurement precision.
Data Source
AI summary
An object is to provide possible web quality estimation for changes in network quality without simulating each combination of network qualities or acquiring operation log information in advance. One aspect of the present invention is a web quality estimation device that estimates web quality which is a processing time of a web content, based on input of a content identifier of the web content and one or more network qualities, and includes: a content characteristic storage unit that stores correspondence information between a content identifier for identifying a web content and one or more content characteristics for a plurality of web contents; and a web quality estimation unit that receives input of a content identifier and one or more network qualities, refers to the corresponding information to acquire one or more content characteristics corresponding to the input content identifier, and estimates web quality from the one or more content characteristics and the input one or more network qualities.