General Internet service assessment by latency including partial measurements

Area of Science:

• Computer Science
• Network Engineering
• Statistics

Background:

• Internet service latency is a critical performance metric.
• Traditional latency assessment methods are often incomplete due to blocked measurements by network infrastructure.
• Underestimation of true service latency can negatively impact user experience and revenue.

Purpose of the Study:

• To develop and validate a novel method for assessing Internet service latency that accounts for blocked measurements.
• To demonstrate the impact of including blocked latency data on service performance evaluation.
• To compare the efficacy of the proposed method against traditional latency assessment techniques.

Main Methods:

• Application of survival theory to process latency data, including blocked measurements.
• Comparative analysis of latency differences between Internet service providers using traditional versus survival latency methods.
• Statistical processing of latency values to quantify underestimation.

Main Results:

• Omission of blocked latency measurements significantly underestimates true Internet service latency.
• The survival latency method revealed a larger latency difference (17 ms) between two providers compared to the traditional approach (9 ms).
• The proposed method provides a more accurate assessment of Internet service performance.

Conclusions:

• The survival latency method offers a more comprehensive and accurate evaluation of Internet service performance.
• Accurate latency assessment is crucial for optimizing online services, particularly in e-commerce and online gaming.
• Incorporating blocked latency data enhances the reliability of service quality metrics.