Internet Resilience in the Asia-Pacific

Member： Christoff VISSER, Romain FONTUGNE, Malte TASHIRO

Category: Exploring

Tags: internet resilience, submarine cables, natural disasters, single points of failure, critical services, internet measurement

  • Background:
    • Recent geophysical events (Hunga Tonga Eruption, Cyclone Gabrielle, etc.) demonstrated the impact of Internet outages due to systems with single points of failure. The Asia-Pacific region is no stranger to natural disasters, and the limited availability of submarine cable infrastructure makes it particularly vulnerable to outages.
  • Purpose:
    • To identify and avoid single points of failure and ensure that critical services in the Asia-Pacific remain available to the communities that need them, even when disaster strikes.
  • Approach:

VULGEO: Vulnerability of the Internet in the South Pacific to Geophysical Events


The Asia-Pacific region is no stranger to natural disasters. The 2022 Hunga Tonga eruption severed the country’s only submarine cable for weeks. Cyclone Gabrielle cut off communities in New Zealand that were assumed to have backup paths. Not only do these events cause physical damage, but they also expose hidden dependencies in the Internet’s infrastructure.


The VULGEO project looks at both the physical infrastructure and the services that keep the Internet working in the region. On the technical side, we use Internet measurement methodologies to identify critical services and their dependencies on offshore hosting services and subsequent submarine cables. On the policy side, we work with network operators and policymakers to translate our findings into actionable guidance.


The project is supported by the Catalyst Seeding Fund from the Royal Society of New Zealand and a grant awarded by the APNIC Foundation. This project is a collaboration between researchers from the IIJ Research Lab, the University of Auckland, Northwestern University, the University of Münster, and the University of Twente and NSRC. Currently, the project is focused on the South Pacific region, but the methodologies and findings are applicable across the Asia-Pacific and beyond.


Our work is roughly divided into the following areas:


Latency Measurements.
We determine where a website is actually hosted, not just where its domain says it lives. Using Internet measurement methodologies, we triangulate the hosting locations of these services. Unlike the rest of the world, the distances between Pacific islands are large enough that latency differences can reveal where traffic really goes, and which services would disappear if a cable fails.


Submarine Cable Criticality.

Which cables would cause the most damage if they failed? Which face the highest physical risk? We use Northwestern University’s Calypso tool to map real traceroute data onto submarine cable paths, then layer geological hazard models on top. We have are modelling the impact across five scenarios across the South Pacific, including a caldera collapse similar to the 2022 Hunga Tonga eruption, impacts of Cyclones and compound events resulting in resource constraints due to multiple disasters happening close to each other.


Classifying Critical Services.

First, we identify the services people need most during emergencies: news, government services, payment providers, and communication platforms. Then we determine where they are hosted: locally or offshore. An example is that a New Zealand emergency management site might have a subdomain resolve to a server in Sydney, while the bare domain points somewhere else entirely. A combination of LLM-based classification and manual review is required to determine which services are critical. Some of our work shows that the region has three distinct Internet cultures: American, Australian/New Zealand, and French, each with their own different patterns of offshore dependency.


Infrastructure and Policy.
Our goal is to translate technical evidence into things people can actually deploy. Recommendations include emergency peering agreements, local hosting proposals, and guidance for funding bodies on where infrastructure investment would most reduce risk.

What happens in the Asia-Pacific previews what will happen elsewhere. There is an increasing trend of consolidation around a single provider, critical services hosted offshore, not peering at the local small IXPs. These failure patterns exist everywhere, not just in the South Pacific. The difference in the South Pacific is the size of the region and the distances between islands.


Presentations

  • Anycast in Underserved Regions — Remi Hendriks (University of Twente). Lightning talk at RIPE 92, Edinburgh, May 2026 (talk)

Blogs

Learn More

The project’s primary website is at vulgeo.com

PAGE TOP