OpenStack is the most widely deployed open-source cloud software in the world. It turns racks of servers into an on-demand cloud of virtual machines, storage, and networks. It runs a large share of the world's private, public, and hybrid clouds, and because it's built on open standards, no single vendor owns your way out. CloudFerro has built its own sovereign cloud on OpenStack since 2014, providing European institutions and companies with the most critical public, private, and hybrid cloud infrastructure.
What is OpenStack?
OpenStack is an open-source platform that pools compute, storage, and networking into a cloud you can run yourself — on your own hardware, in your own data centre, or through a provider that runs it for you. It turns physical servers into on-demand virtual machines, block and object storage, and virtual networks, controlled through APIs, a command line, or a web dashboard.
OpenStack includes four core capability areas:
- Compute
- Storage
- Networking
- Identity
OpenStack is an industry standard with more than 55 million compute cores run on it in production worldwide as of 2025, across telcos, banks, research institutions, and European public clouds — with mega-users such as CERN, Walmart, and Workday each running over a million cores.
OpenStack vs. VMware
OpenStack is open-source cloud infrastructure you build a cloud with; VMware is proprietary virtualisation licensed per host inside its own ecosystem. VMware gives you a polished, tightly integrated product while OpenStack gives you a modular, API-driven cloud with no licence fee and no single vendor controlling the pricing — a packaged product you accept lock-in for, versus a portable stack you control.
OpenStack vs. hyperscalers
A hyperscaler such as AWS or Azure is a proprietary public cloud you rent from one a US-based provider. OpenStack is the open-source software that you — or a European provider — use to run a public, private, or hybrid cloud you control. The interfaces look similar. The difference is who owns the platform.
Why do open standards matter for sovereignty?
Open standards and open source mean no single vendor controls the code, the pricing, or your way out — and that independence is a precondition for a sovereign cloud. This works on two levels:
- No lock-in: open code and open APIs let you move workloads between providers, reuse existing tools, and pull back on-premises without re-platforming penalties or exit fees.
- Auditability and independence: open-source code can be inspected and run by any operator. There's no foreign-controlled control plane, no closed roadmap, and no vendor kill-switch.
Real sovereignty needs both open standards and an EU operator that owns the stack and the hardware.
Open source vs. proprietary cloud
Proprietary platforms (VMware, or a hyperscaler's managed stack) give you polish and one vendor to call, but the licence terms, roadmap, and pricing sit with a single company, often outside Europe. Open source on a managed European provider gives you portable, inspectable, lock-in-free infrastructure — at the cost of needing an operator with real expertise to run it well.
Proprietary suits teams that want a packaged product while open source suits regulated, public-sector, and sovereignty-sensitive workloads.
What is OpenStack used for?
OpenStack gives you a full set of cloud services from one framework — virtual machines, bare metal, containers, block and object storage, backups, databases, and virtual networking — and runs public clouds, private clouds, and hybrids of the two.
- Public and private cloud: providers and enterprises run OpenStack to offer, or self-host, on-demand infrastructure they fully control.
- Telecoms / NFV: carriers run network functions as software on OpenStack — a major share of production deployments, and a big reason it scaled to tens of millions of cores.
- Containers at scale: OpenStack provisions the VMs, networks, and storage that Kubernetes clusters run on.
- Research, AI: heavy compute and data workloads kept under one jurisdiction.
OpenStack and Kubernetes
OpenStack provides the infrastructure, and Kubernetes runs the containers on top of it. OpenStack hands out VMs, networks, block storage, and bare metal; Kubernetes schedules containers across that capacity.
They connect both ways — OpenStack can provision Kubernetes clusters through Magnum, and Kubernetes can request OpenStack resources such as block volumes and load balancers through its cloud-provider integration. Many teams run Kubernetes on OpenStack to get both layers on one open, sovereign stack, with no proprietary cloud underneath.
Key OpenStack concepts
A few terms you'll meet around OpenStack:
- Open standards: published, shared interfaces any vendor can implement — what keeps workloads portable.
- IaaS (infrastructure as a service): raw compute, storage, and networking delivered on demand — the layer OpenStack provides.
- Multi-tenancy (projects / tenants): isolated spaces so many teams or customers share one cloud without seeing each other's resources.
- Control plane: the OpenStack services that schedule and manage resources, as distinct from the workloads running on them.
- API endpoint & service catalogue: the address each service publishes, listed in one catalogue so tools can find them.
How does OpenStack work?
OpenStack is built from separate services that each run one part of the cloud and talk to each other through public APIs.
- Every service authenticates through one shared identity service (Keystone), and services call each other through public APIs, passing work internally through a message broker.
- State lives in a replicated database, while the working processes are stateless — so you scale them horizontally for capacity and high availability.
- Users reach the cloud three ways: a REST API, a command-line interface, or the Horizon web dashboard.
- New services can be added without redesigning the rest, which is how OpenStack grew from compute and storage to bare metal and containers.
What are the main OpenStack components?
OpenStack is a modular framework with around 34 official components, grouped into four categories:
- OpenStack Services (the core cloud services)
- Operations tooling
- Add-ons to services
- Integration enablers.
Most users only ever touch a handful of the core services — the ones you'll meet most often:
- Keystone — identity: authentication, authorisation, and the service catalogue every other service registers with.
- Nova — compute: provisions and manages virtual machines.
- Neutron — networking: virtual networks, routers, floating IPs, and security groups.
- Cinder — block storage: persistent volumes you attach to VMs.
- Glance — image service: stores and serves the OS images that VMs boot from.
- Horizon — dashboard: the web console for the whole cloud.
- Ironic — bare metal: provisions physical servers the same way as VMs.
- Designate — DNS-as-a-service.
- Octavia — load balancing: managing virtual machines, containers, or bare metal servers.
- Barbican — secret data: secure storage, provisioning and management passwords, encryption keys, X.509 Certificates and raw binary data.
A provider can also swap the backend behind these services — for example, CloudFerro uses Ceph behind its storage.
How does CloudFerro build its cloud on OpenStack?
CloudFerro has built its clouds on OpenStack since 2014, and runs its public and private cloud on it across Europe today.
- Over a decade of OpenStack: from the first "Juno"-based deployment in 2014 to current releases, CloudFerro's engineers have integrated, operated, scaled, and developed OpenStack clouds, with a dedicated OpenStack team available round-the-clock.
- The stack it sits in: OpenStack for compute and orchestration, Ceph for storage, and Kubernetes for containers, plus CloudFerro's own tenant-management, billing, and provisioning layer integrated with OpenStack.
- Scale: installations serving thousands of users and managing thousands of servers, terabytes of RAM, and petabytes of storage across Europe.
- Open interfaces: S3-compatible and standard APIs, so existing workloads and tools migrate in without a rewrite.
Why choose CloudFerro's open-source cloud?
- Sovereign at every layer: an EU-owned company, under EU jurisdiction, in EU data centres, on an open-source stack you can audit and leave — no US CLOUD Act exposure, and no proprietary control plane you can't see into.
- No vendor lock-in, easy in, easy out: Open, standard OpenStack APIs mean your workloads, tooling, and automation aren't tied to CloudFerro. Migrating in is easy and cheap. You stay because the cloud works, not because you're stuck.
- Own stack, own hardware: CloudFerro designs, builds, and operates its own OpenStack-based cloud, with Ceph and Kubernetes, on its own hardware. There are no non-EU subcontractors anywhere in the chain.
- OpenStack expertise since 2014: over a decade integrating, scaling, and running OpenStack in production, with a team on it round-the-clock.
- Certified and cleared: ISO 9001 and ISO/IEC 27001:2022 (with the 27017 and 27018 control sets), BSI C5, and clearance for classified workloads.
- Proven on critical programmes: the same open-source cloud runs the Copernicus Data Space Ecosystem and 100+ PB of Earth-observation data for institutions such as ESA, the European Commission, and EUMETSAT.
How is this different from a hyperscaler or a VMware-based cloud?
Most "EU cloud" offers still run on proprietary software, foreign-owned control planes, or non-EU subcontractors underneath. CloudFerro built its own cloud on open source and owns it end to end — from the OpenStack services down to who racks the hardware. You get an open, portable stack with no vendor lock-in, run under full EU governance, and priced without proprietary licence spirals.
Frequently asked questions
Build on an open, sovereign European cloud
Run your workloads on a cloud built end-to-end on open source and owned in Europe. Talk to CloudFerro's OpenStack experts about your migration, or a private cloud of your own.