Smartphone-based crowdsourcing for estimating the bottleneck capacity in wireless networks

TL;DR

This paper introduces SmartProbe, a smartphone-based crowdsourcing system that efficiently estimates network bottleneck capacity and compares mobile broadband performance over large areas using georeferenced data.

Contribution

The paper presents a novel, bandwidth-efficient crowdsourcing system for network measurement that supports large user bases and provides georeferenced performance insights.

Findings

SmartProbe is more bandwidth-efficient than similar systems.

Supports a larger number of users for network measurement.

Provides valuable georeferenced data for operator performance comparison.

Abstract

Crowdsourcing enables the fine-grained characterization and performance evaluation of today's large-scale networks using the power of the masses and distributed intelligence. This paper presents SmartProbe, a system that assesses the bottleneck capacity of Internet paths using smartphones, from a mobile crowdsourcing perspective. With SmartProbe measurement activities are more bandwidth efficient compared to similar systems, and a larger number of users can be supported. An application based on SmartProbe is also presented: georeferenced measurements are mapped and used to compare the performance of mobile broadband operators in wide areas. Results from one year of operation are included.