When you send a message or stream a video, you rarely think about where that data actually goes.
Yet much of it travels through thin glass fibres on the seabed, connecting continents. For Lynn van der Horst, a master’s student in Internet Science & Technology at UT, those invisible connections are what make the internet both fascinating and fragile. She is currently doing her graduation project at the RIPE NCC, an organisation that helps coordinate and support the technical infrastructure of the internet in Europe and the Middle East. She studies how the Internet behaves when something goes wrong.
Measuring the internet’s health
RIPE NCC (amongst others) allocates both IP addresses (the numbers that identify every device online) and AS numbers in its service region, but also operates RIPE Atlas, a network of thousands of small measuring devices across the world. These devices send tiny test signals, such as pings, to track how data moves through the internet.
Lynn’s research focuses on spotting signs of congestion in that data: moments when certain routes become overloaded and traffic slows down. “It’s similar to a traffic jam,” she explains. “Too many data packets try to pass through the same path and you start to see delays, or even packet loss.” By analysing patterns in the timing of these signals, she can detect where and when the internet experiences slowdowns. Occasionally, those patterns reveal something larger, such as a damaged undersea cable.
Disruptions in global connectivity
Earlier this year, several internet cables in the Red Sea broke, causing disruptions in global connectivity. For Lynn, it was a chance to apply her methods to real data. “When a cable fails, traffic has to be rerouted. You can see that in the measurements: suddenly data takes different, often longer paths,” she says. These incidents highlight how dependent the modern world is on physical infrastructure. “We often think of the internet as something abstract,” Lynn says. “But it relies on a relatively small number of physical cables. Understanding what happens when something goes wrong helps us assess how resilient the network really is.”
Early detection is essential
Detecting slowdowns or outages quickly helps researchers and the wider internet community understand how and why disruptions happen. While routing adjustments are often made automatically by network protocols such as BGP, and cable operators are usually the first to know when their infrastructure fails, external measurements like those from RIPE Atlas provide valuable insight into the broader effects of an outage. They make visible how traffic patterns change and how resilient the network really is.
Understanding these patterns doesn’t prevent outages, but it helps to reveal their consequences and severity. Because the internet is decentralised and much of its physical structure is not publicly known, predicting how failures will spread remains extremely difficult. Yet our daily lives depend on this hidden infrastructure, “just working.” Lynn’s research helps shed light on what happens underneath when it doesn’t.
Cybersecurity
Cybersecurity is often associated with passwords, viruses or hackers, but resilience plays an equally important role. “My research doesn’t directly prevent attacks,” Lynn says, “but it helps us understand how the internet responds to disruptions.” That insight can make networks more reliable. “If we want to protect the internet, we must be aware of its vulnerabilities and what would happen if they are in any way breached. Whether it’s a technical fault or something intentional,” she explains.
Lynn’s interest in the topic comes from a simple curiosity about how the internet works. “It’s easy to take for granted that everything is connected and available at all times,” she says. “But underneath it all is a very complex system that depends on cooperation, technology and a good bit of luck.”
