Throughput Measurement Latency Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Key Performance Indicators (KPIs) for data communications in cellular radio systems fail to accurately represent end-user experienced throughput, particularly in high-bandwidth systems, due to inclusion of latency periods and vendor-specific measurements, making it difficult to compare and benchmark performance across different systems and vendors.
Innovation Solution
A method to calculate data transfer throughput by compensating for latency time periods, focusing on the time when actual data is transferred, and providing performance measures that are independent of the total data size, using standardized interfaces to ensure objective benchmarking across systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If mean data rate is used as a KPI to measure throughput, then it is easy to implement and calculate, but it does not accurately represent end-user perceived quality because it includes periods of inactivity
Solution Approach 1:
The patent extracts and removes the latency period from the throughput calculation. It identifies the initial period where no data is transmitted despite data being available in the buffer, and excludes this period from the throughput measurement. This is achieved by detecting when data arrives in an empty transmit buffer and calculating the time until the first transmission, then subtracting this latency from the total measurement period to provide an accurate end-user perceived throughput.
2Measurement precision
If throughput is measured only when data is in the transmit buffer, then it gets closer to end-user perception, but it does not account for the initial latency period before transmission starts
Solution Approach 1:
The patent extracts and removes the latency period from the throughput calculation. It identifies the initial period where no data is transmitted despite data being available in the buffer, and excludes this period from the throughput measurement. This is achieved by detecting when data arrives in an empty transmit buffer and calculating the time until the first transmission, then subtracting this latency from the total measurement period to provide an accurate end-user perceived throughput.
Solution Approach 2:
The patent performs preliminary detection and measurement of the latency period. It monitors the transmit buffer to detect when data arrives and when transmission starts, calculating the latency time in advance. This preliminary measurement of the latency period enables the subsequent exclusion of this time from the throughput calculation, ensuring accurate end-user perceived throughput measurement.
3Adaptability or versatility
If different KPI measures are used in different systems and vendors, then each vendor can optimize for their specific system, but it makes it difficult to compare and benchmark performance across different systems and vendors
Solution Approach 1:
The patent defines a universal throughput measurement method that can be applied across different wireless communication systems and vendors. The approach of measuring throughput during active transmission periods and excluding latency periods is vendor-agnostic and system-independent. This universal method enables fair comparison and benchmarking of performance across different systems (GSM, WCDMA, HSPA, LTE, WiMAX) and vendors while maintaining the ability to capture end-user perceived quality.
Data Source
AI summary
The technology in this application determines the throughput over a data communication substantially as experienced by a user receiving the data. The determined throughput is substantially independent of the total amount of data transferred over the connection. This “active” throughput measure is divided into two parts to provide two performance measures: transfer throughput and latency. Latency corresponds to a time period needed to accomplish one or more functions associated with the data connection, like setting up and taking down the connection, during which payload type data to be transferred to an end user is not transferred. By identifying and removing the initial latency from the throughput determination for the data connection, the determined transfer throughput provides a more accurate measure of the throughput actually experienced by the end user.