Thomas Ornell

Most teams talk about cloud sovereignty and provider-exit. Almost nobody has actually built it and can tell you what it costs. This is the real story of building a US-independent fallback environment for a European telematics platform running half a million connected assets: how we selected the solutions and why, and the genuinely hard parts of making sure the core can run outside the major cloud providers at all. Pain and all.

To be clear, we're not leaving the US owned cloud unless we have to. The point is to be ready if we're ever forced. Everything is built from open source parts you can reproduce yourself, Kafka, HiveMQ, MinIO,OpenBao, GitLab, Harbor, and ArgoCD, so there's no vendor lock and no "only works at our budget."

And it's an honest account of an ongoing journey, including the part most architecture talks skip: resilience turned out to be an organisational property as much as a technical one. One person quietly owning 13+ systems was a bigger DR risk than any single component we designed around. You'll leave with the solution-selection reasoning, a concrete map of where running outside the big providers actually bites, and the successes, failures, and hidden traps from a project that's still very much underway.