MAnycast Reloaded: a Tool for an Open, Fast, Responsible and Efficient Daily Anycast Census

2025-03-26

AI Summary

Existing IP anycast census tools face a precision-efficiency trade-off: iGreedy achieves high accuracy but suffers from low efficiency and high cost, whereas MAnycast2 is efficient yet insufficiently accurate. This paper introduces MAnycast Reloaded—the first open-source, active measurement framework that unifies the strengths of both approaches. It proposes a novel two-stage pipeline: “accuracy-oriented” probing followed by “speed-optimized” refinement. The framework supports multi-protocol probing (UDP, TCP, IPv6), distributed global deployment, RIPE Atlas–based latency calibration, and community-driven validation. Evaluated on a 32-node testbed and real production networks, it achieves >90% precision and accuracy in anycast prefix identification, with over 60% of results independently confirmed by network operators. Daily updated census data has been publicly available for sustained operation. This work establishes the first常态化 (routine) anycast census solution that simultaneously delivers high accuracy, low overhead, full protocol support, and verifiability.

Technology Category

• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Search and Optimization: Distributed Search
• Machine Learning: Active Learning

Application Category

• Responsible Web: Measurement, analysis, and circumvention of Web censorship
• Web Mining and Content Analysis: Web measurements
• Security and Privacy: Large-scale security measurements

Abstract

IP anycast is a widely adopted technique in which an address is replicated at multiple locations, to, e.g., reduce latency and enhance resilience. Due to anycast's crucial role on the modern Internet, earlier research introduced tools to perform anycast censuses. The first, iGreedy, uses latency measurements from geographically dispersed locations to map anycast deployments. The second, MAnycast2, uses anycast to perform a census of other anycast networks. MAnycast2's advantage is speed, performing an Internet-wide census in 3 hours, but it suffers from problems with accuracy and precision. Inversely, iGreedy is highly accurate but much slower. On top of that, iGreedy has a much higher probing cost. In this paper we address the shortcomings of both systems and present MAnycast Reloaded (MAnycastR). Taking MAnycast2 as a basis, we completely redesign its measurement pipeline, and add support for distributed probing, additional protocols (UDP, TCP and IPv6) and latency measurements similar to iGreedy. We validate MAnycastR on an anycast testbed with 32 globally distributed nodes, compare against an external anycast production deployment and extensive latency measurements with RIPE Atlas, and cross-check over 60% of detected anycast prefixes against operator ground truth. This shows that MAnycastR achieves high accuracy and precision. We make continual daily MAnycastR censuses available to the community and release the source code of the tool under a permissive open source license.

Problem

• Improves accuracy and speed of anycast census tools
• Supports multiple protocols and distributed probing
• Validates anycast detection with high precision

Innovation

• Redesigned measurement pipeline for efficiency
• Supports UDP, TCP, IPv6 protocols
• Integrates latency measurements for accuracy