For most independent cloud providers, the strategy conversation ends in the same place. You can try to become a hyperscaler, which takes capital only a handful of global companies can raise. Or you can resell one and call it a partnership. Either way, the customer relationship, the pricing and the roadmap drift toward whoever owns the platform.
Governments hear a version of the same offer: take a region of a global platform and call it sovereign. Data centre operators hear it as a hyperscaler tenancy in their halls. The shape is always the same trade: you can have a cloud that works, or a cloud that’s yours.
Federation is the way out of that trade. Not a new kind of hyperscaler, but independent clouds, each owned and run locally, connected by explicit agreement.
What a federated cloud network is
A federated cloud network is a group of independently owned clouds that share a common platform, common standards and a common way of doing business with one another. Each member keeps its own infrastructure, customers, pricing and operations. What they share is the ability to offer services to each other’s customers.
The closest analogy is an airline alliance. Each airline keeps its brand, its fleet, its crews and its passengers. Together they sell journeys none of them could fly alone. Nobody thinks a member airline has given up its independence by joining. It has extended its network without buying more aircraft.
Why the centralised model is being questioned
Centralised public cloud delivered enormous value, and nothing here pretends otherwise. But concentration brings its own problems:
- Dependency — on one provider’s pricing, roadmap and terms.
- Sovereignty — residency, jurisdiction and local governance that a control plane in another country cannot fully answer.
- Economic leakage — cloud spend flowing out of local technology economies instead of building them.
- Distance — support, expertise and accountability that sit far from the customer.
Each of these pushes toward local clouds. The objection has always been that local clouds are too small to compete. Federation is the answer to the objection.
Five principles that make it work
Independence: every member owns its infrastructure and its customer relationships. Interoperability: services behave consistently across members because they run on the same platform. Shared standards: members follow common operational, security and governance practices, so a customer can trust a service whichever cloud runs it. Collaboration: opportunities, supply and specialisms are shared deliberately, not by default. Sovereignty: each member keeps control of its own data and its own compliance obligations.
The last principle is the easiest to promise and the hardest to keep. So it deserves precision.
What stays home, and what travels by agreement
Federation is often sold as a network diagram. The only question that matters is what crosses the boundary between clouds, and what never does.
Stays in your cloud: customer accounts and contracts, customer data and workloads, your pricing and margins, your infrastructure and operations.
Travels by agreement: the catalogue items you choose to share, service orders and their fulfilment, usage records for settlement, and discovery of what other clouds offer.
It is worth being honest about the edge case. If you choose to burst, fail over or run disaster recovery into a partner cloud, the workload and its data run there — because you configured it, under an agreement. Nothing places itself. Customer data never moves automatically, and customer records have no path out at all.
Federation moves commerce by default, and workloads only by configuration.
Connect. Publish. Sell. Settle.
In practice federation is a business process, not an integration project. Each cloud’s Oasis Core connects to the Oasis Federation Core, run by CloudHost, which brokers every exchange; no identity store is shared. Each publishes the catalogue items it is willing to share, and with whom. Customers order a federated service from their own provider’s portal, on their home account. Usage is metered, and the purchase is brokered through the Federation Core: the delivering partner invoices CloudHost, and CloudHost invoices the requesting partner, who bills its own customer.
If a partner leaves, its published services stop being orderable and running ones are handled under the agreement. Your own cloud, customers and catalogue continue unchanged. Federation is participation, not dependency.
What each participant gains
Cloud operators and MSPs gain reach and supply. They can sell services they do not run themselves — GPU capacity, disaster recovery sites, ISV software, managed services — under their own brand, and reach customers in markets where they never built a presence.
Data centre operators can move from landlord to provider, and give their tenants a way into a larger ecosystem instead of competing with them.
Governments get national clouds that keep data, identity and operations at home and still cooperate across borders, without surrendering any of it.
AI programmes gain access to specialised capacity that not every member needs to build, with local governance intact.
And the marketplace compounds. Every new member adds services every other member can sell, so the network becomes more valuable as it grows.
The economics have to match the principle
A federation that takes a share of every member’s growth is centralisation under another name. The commercial model has to reward participation without taxing it. On Oasis Stack, federation is a flat participation fee, the platform is a flat fee per deployment banded by scale, and what each operator charges its own customers stays its own.
Oasis Stack is built around this model. Each cloud’s Oasis Core holds its local federation connection, and the Oasis Federation Core carries the four moves, from publishing to settlement. CloudSouq is the storefront through which each cloud sells, under its own brand and at its own margin. And CloudHost runs its own public cloud, YallaCloud, on the same platform.
Sovereign. Not isolated.
For a long time, cloud growth was driven by centralisation. The next phase will be shaped by clouds that are local enough to be trusted and connected enough to compete. Independent providers, governments, telecom operators, MSPs and data centres are reaching the same conclusion: they can do more together than alone, without giving up what makes them independent.
That is the belief federation makes practical. Sovereignty should not cost you scale.