Designing for failure: blast-radius thinking in mission-critical networks

Tuesday, November 17, 2026 at 3:30 PM to 4:00 PM · 30 min. (Europe/Luxembourg)
Room F01
LUNOG9

Information

Most major outages of recent years were not caused by hardware failing, but by changes and software behaving in ways nobody had tested, and they became headlines because the failure domain was far larger than anyone realized. This talk takes the view that resilience is not only about redundancy but also about limiting how far any single failure can travel. We walk through concrete design patterns that keep the blast radius small: splitting control-plane and failure domains, keeping the underlay and overlay independent, staged and hitless upgrade strategies, validating every change against a digital twin before it touches production, and knowing in advance exactly what you lose when a site, a route server or a management plane goes dark. Examples are drawn from real deployments in critical networks, where downtime has a direct impact on people.

Join the event!

See all the content and easy-to-use features by logging in or registering!