Project Overview
The RABBITS (Reproducible Assessment of BroadBand Internet Topology and Speed) project aims to address the critical gaps in existing broadband performance measurement tools by developing a comprehensive toolkit for Reproducible Assessment of BroadBand Internet Topology and Speed. This toolkit will enable consistent and reliable broadband performance measurements across platforms, ensuring accurate and reproducible data that is essential for both scientific research and public policy.
Key Challenges Addressed:
-
Server Deployment and Selection: Current speed test platforms suffer from uneven server deployment and suboptimal server selection, which can distort measurement results. The RABBITS toolkit will enable the discovery and characterization of speed test servers to ensure more accurate measurements.
-
Inconsistent Test Parameters: Different platforms use proprietary and varying test parameters, making it impossible to compare results across platforms. RABBITS will standardize these parameters to facilitate reproducibility and comparability in broadband measurements.
Figure 1: Placement of test servers (e.g., on-net, transit ISPs, CDNs) affects the measured network paths, and sometimes may not align with a user’s measurement goals. Lack of access to nearby servers (e.g., for User R) may lead to confusing test results. Measurement traffic sent to a remote on-net server in a different region, or to off-net test servers, will impede the accuracy of last-mile measurements. In contrast, user M with an onnet server in the same region will have low latency and interference with last-mile speed tests. High test server density will also enable users to crossvalidate performance measurements using servers in different types of networks.
Modules of RABBITS:
-
RABBITS-Mapper: This module will discover and map the locations of test servers hosted by multiple platforms. It will periodically collect geographical and topological data, using strategically deployed vantage points (VPs) in CAIDA’s Ark, RIPE Atlas, and EdgeNet to characterize the paths and latency from end-users to speed test servers.
-
RABBITS-Perf: This module will allow researchers to control test parameters, such as the number and size of concurrent measurement flows, and select specific test servers. It will also standardize the format of measurement data to ensure consistency and reproducibility across different platforms.
Supported Platforms:
The RABBITS project supports a variety of speed test platforms, each with unique characteristics and deployment scales. Below is a summary of the platforms that RABBITS will support:
| Platform | # of Servers (IPv4) | # of Servers (IPv6) | # of Countries | # of ASes | Type | Public Server List |
|---|---|---|---|---|---|---|
| Ookla | 18,997 | 2,073 | 216 | 8,901 | Unicast | No |
| M-Lab | 181 | 181 | 36 | 43 | Unicast | Yes |
| Comcast | 139 | 139 | 1 | 1 | Unicast | No |
| CloudHarmony - AWS | 23 | 0 | 17 | 1 | Unicast Cloud | No |
| CloudHarmony - Azure | 36 | 0 | 17 | 1 | Unicast Cloud | No |
| CloudHarmony - GCP | 24 | 0 | 17 | 1 | Unicast Cloud | No |
| Fast.com | 300 | Unknown | 28 | 6 | Unicast CDN | No |
| Speedof.me | 127 | 127 | 32 | 1 | Anycast CDN | No |
| Cloudflare | Unknown | Unknown | >100 | 1 | Anycast CDN | No |
| FCC | 133 | 133 | 1 | 15 | Unicast | No |
Policy Engagement
The RABBITS project will support critical broadband policies, including the FCC Broadband Consumer Labels, the new FCC broadband maps, and NIST’s BEAD program, by providing accurate data on network performance. This data will help regulators and policymakers ensure fair broadband access, identify underserved communities, and validate the performance claims made by ISPs. Additionally, ISPs can use RABBITS’ insights to optimize their network infrastructure and improve service delivery.
Broader Impacts
The RABBITS project will enhance scientific understanding of speed test platforms and enable reproducible broadband performance experiments across platforms. The data and tools developed will benefit consumers, industry, and regulators by providing reliable insights into network performance, supporting policy decisions, and guiding ISP infrastructure improvements. The project will also contribute to education by creating accessible materials and engaging undergraduate students, particularly from underrepresented groups, through the Department of Computer Science and Engineering (CSE) and Halıcıoğlu Data Science Institute (HDSI) at UCSD.
Data Providers
Funding Support
This material is based on research sponsored by the National Science Foundation (NSF) grant CNS-2323219. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of NSF.