Tesco is migrating approximately 40,000 server workloads from VMware to an open-source platform following Broadcom's decision to discontinue support. The migration is proceeding at an exceptional pace, requiring the retailer to adapt or develop new tools for data protection, as existing solutions like Veeam and Zerto were incompatible with the new environment.
The migration process involves nine functional axes, categorized into features that work immediately, those requiring manual configuration, and those necessitating new operational procedures. While virtual machines can be moved via conversion tools, the surrounding infrastructure—including backup systems, microsegmentation, and disaster recovery plans—must be rebuilt.
Backup systems in open-source platforms often rely on Changed Block Tracking (CBT) mechanisms, which vary by platform. For instance, Proxmox VE uses QEMU dirty bitmaps, while oVirt-based solutions utilize checkpoints. These differences impact backup efficiency and data consistency, requiring careful planning and testing of guest agents for application-level integrity.
Orchestration layers, such as OpenStack or oVirt, differ from hypervisors like KVM. OpenStack has seen significant adoption, with production deployments reaching 55 million cores by October 2025. Managing infrastructure as code is possible via APIs and Terraform, though some custom forks require proprietary development.
Advanced features like DRS-style auto-balancing are evolving. OpenStack uses the Watcher component, which integrated Prometheus metrics in the April 2025 Epoxy 2025.1 release. Meanwhile, Proxmox VE lacks a native balancer, necessitating custom scripts for large-scale clusters. Microsegmentation is typically handled via OVN, though L7 application identification may require additional components.
