Speed Test Analysis

Overview

Our speed test analysis research examines the methodologies, accuracy, and best practices for measuring Internet performance. This work is critical for informing broadband policy, as speed test data is increasingly used to allocate billions of dollars in infrastructure investments.

Research Areas

Comparative Analysis

We conducted the first comprehensive comparison of popular speed test platforms (Ookla Speedtest and M-Lab NDT7) in both controlled lab settings and real-world deployments. Our findings reveal significant differences in how these tools measure performance.

Best Practices

Based on analysis from controlled experiments and a six-month deployment across 77 households in Chicago, we developed recommendations for gathering and analyzing crowdsourced speed test measurements to support broadband investment allocation.

Latency Under Load

Beyond simple throughput measurements, we developed techniques to measure “operational latency”—latency under load conditions that better reflect real-world application performance.

Key Findings

• Speed test results vary significantly based on test methodology
• Client device, access medium (WiFi vs. wired), and server selection all impact results
• Proper methodology is essential for policy-relevant measurements

Publications

A Comparative Analysis of Ookla Speedtest and Measurement Labs Network Diagnostic Test (NDT7) Kyle MacMillan, Tarun Mangla, James Saxon, Nicole P. Marwell, Nick Feamster ACM SIGMETRICS, 2023

Best Practices for Collecting Speed Test Data Kyle MacMillan, Tarun Mangla, Nick Feamster, Nicole Marwell TPRC, 2022

Measuring Internet Speed: Current Challenges and Future Recommendations Nick Feamster, Jason Livingood Communications of the ACM (CACM), 2020

Beyond Speed Test: Measuring Latency Under Load Across Different Speed Tiers Kyle MacMillan, Nick Feamster IAB Workshop on Measuring Network Quality for End Users, 2021