https://tube.tuxis.nl/videos/watch/bc03ba12-c17f-4bfb-93c3-e1c14e1409c6 Alexandre Derumier is a French systems architect, open-source contributor, and virtualization specialist best known for his long-standing work around the Proxmox VE ecosystem and his technical leadership, working at Cyllene, French IT infrastructure and managed services company. With more than a decade of experience in infrastructure engineering, storage systems, and high-availability architectures, Alexandre has specialized in designing and optimizing enterprise virtualization platforms for web companies, e-commerce businesses, and media organizations. Alexandre is also widely recognized within the Proxmox open-source community. He contributes to technical discussions, feature proposals, and patches on the Proxmox development mailing lists, particularly around networking, storage, SDN, IPAM integration, clustering, and virtualization management improvements. Beyond development work, he is known for training and knowledge sharing. Alexandre Derumier is one of the official Proxmox trainers in France and has delivered specialized courses covering Proxmox VE deployment, clustering, shared storage, and enterprise virtualization operations. He has also represented the Proxmox and open-source virtualization community at technical events such as FOSDEM in Brussels, where he participated alongside other infrastructure and virtualization experts to showcase Proxmox technologies and enterprise deployments. From 100 to 10,000 VMs: 15 Years of Proxmox in Production at Scale Over the last 15 years, we have built, operated, and continuously evolved large-scale production infrastructures based on Proxmox VE — growing from a few hundred virtual machines to more than 10,000 running workloads. In this talk, Alexandre Derumier will share the real-world journey behind operating Proxmox at scale: the architectural choices we made, the technical challenges we faced, and the lessons learned through years of production experience. We will cover how the platform evolved over time, from early KVM and OpenVZ deployments to modern hyperconverged infrastructures with Ceph, SDN, automation, and high availability. The session will also explore the contributions we made to the Proxmox ecosystem, including features, bug fixes, scalability improvements, and operational tooling developed from real production needs. Expect practical feedback on performance tuning, clustering limits, storage design, networking, migrations, failures, upgrades, and scaling strategies — with both successes and mistakes from the field. This talk is aimed at infrastructure engineers, hosting providers, cloud architects, and anyone interested in running open-source virtualization platforms reliably at large scale. Fri, 25 Sep 2026 21:08:42 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://tube.tuxis.nl https://tube.tuxis.nl/lazy-static/avatars/ae66bff6-6fa0-4e99-ab39-86bc7f985c17.png https://tube.tuxis.nl/videos/watch/bc03ba12-c17f-4bfb-93c3-e1c14e1409c6 All rights reserved, unless otherwise specified in the terms specified at https://tube.tuxis.nl/about and potential licenses granted by each content's rightholder.