TURBOTEST: Learning When Less is Enough through Early Termination of Internet Speed Tests

2025-10-24

AI Summary

This paper addresses redundant data transmission in Internet speed testing by proposing a precision-aware early termination mechanism. Methodologically, it introduces a two-stage learning framework: the first stage employs a regression model to predict instantaneous throughput using transport-layer features (e.g., RTT, retransmission rate, congestion window); the second stage deploys a classifier to dynamically decide whether to terminate the test, supporting configurable accuracy tolerance thresholds and adaptive fallback under high-throughput volatility. Evaluated on 173,000 real-world M-Lab NDT measurements, the approach reduces upload traffic by 2–4× compared to BBR-based signaling while lowering median absolute error—significantly improving the accuracy-overhead trade-off. The core contribution lies in decoupling throughput prediction from termination decision-making, enabling robust, accuracy-guaranteed early stopping with tunable precision.

Technology Category

• Machine Learning: Online Learning & Bandits
• Search and Optimization: Algorithm Configuration
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Security and Privacy: Large-scale security measurements
• Web Mining and Content Analysis: Web measurements

Abstract

Internet speed tests are indispensable for users, ISPs, and policymakers, but their static flooding-based design imposes growing costs: a single high-speed test can transfer hundreds of megabytes, and collectively, platforms like Ookla, M-Lab, and Fast.com generate petabytes of traffic each month. Reducing this burden requires deciding when a test can be stopped early without sacrificing accuracy. We frame this as an optimal stopping problem and show that existing heuristics-static thresholds, BBR pipe-full signals, or throughput stability rules from Fast.com and FastBTS-capture only a narrow portion of the achievable accuracy-savings trade-off. This paper introduces TURBOTEST, a systematic framework for speed test termination that sits atop existing platforms. The key idea is to decouple throughput prediction (Stage 1) from test termination (Stage 2): Stage 1 trains a regressor to estimate final throughput from partial measurements, while Stage 2 trains a classifier to decide when sufficient evidence has accumulated to stop. Leveraging richer transport-level features (RTT, retransmissions, congestion window) alongside throughput, TURBOTEST exposes a single tunable parameter for accuracy tolerance and includes a fallback mechanism for high-variability cases. Evaluation on 173,000 M-Lab NDT speed tests (2024-2025) shows that TURBOTEST achieves nearly 2-4x higher data savings than an approach based on BBR signals while reducing median error. These results demonstrate that adaptive ML-based termination can deliver accurate, efficient, and deployable speed tests at scale.

Problem

• Reducing data transfer costs in internet speed tests
• Determining optimal early termination for accuracy preservation
• Decoupling throughput prediction from test termination decisions

Innovation

• Two-stage ML framework for throughput prediction
• Uses transport-level features for termination decisions
• Single tunable parameter controls accuracy-savings tradeoff