OCI Dedicated Region: The Full Cloud in Your Data Center

Most cloud offerings that claim to bring the cloud to you bring a slice of it: a rack of database machines, an appliance, an outpost that depends on a parent region hundreds of miles away. OCI Dedicated Region is a different proposition. Oracle ships, installs, and remotely operates an entire cloud region, with the full public service catalog, inside a facility you own or control. To your developers it looks exactly like any other OCI region: the same console, the same APIs, the same Terraform providers, the same SLAs, the same per unit prices. To your regulator it looks like something far more interesting: a hyperscale cloud that never leaves the building. It is the most extreme answer on the sovereignty spectrum we map in our complete guide to OCI compliance and sovereignty, and like most extreme answers, it is exactly right for a specific set of buyers and expensively wrong for everyone else.

This article covers what Dedicated Region actually is, who it is for, the division of responsibilities between you and Oracle, the commercial shape, how it differs from the single service Cloud at Customer offerings, how identity and realms work, what operations look like once it is running, and the risks that surface after the contract is signed. It ends with an eight step evaluation sequence you can run before talking to Oracle.

What Dedicated Region actually is

A Dedicated Region is a full OCI region, the same physical and logical architecture Oracle deploys in its own public regions, built inside the customer's data center. That means the complete service catalog: compute in all its shapes, block and object storage, the full networking stack, Autonomous Database and Exadata services, the Kubernetes engine, analytics, integration, security services, and the Oracle SaaS applications that run on OCI. Not a subset chosen at contract time, the catalog, with new services arriving as they roll out across the realm.

Oracle owns the hardware, installs it, and operates it remotely through the same control plane and the same operations organization that runs its public regions. Patching, capacity management inside the contracted footprint, hardware replacement, and service updates are Oracle's job, performed without your teams touching the gear. The region behaves as a first class member of OCI: it appears in your tenancy, your existing automation targets it by changing a region parameter, and consumption is metered in the same Universal Credits units at the same published rates as public regions. The cloud experience is preserved end to end; the only thing that moved is the physical location.

That last point is why the offering has held attention since it launched. Competing on premises offerings tend to make architectural compromises: a reduced catalog, dependence on a remote parent region for the control plane, or different pricing. Dedicated Region's pitch is that there is no second class citizenship. If a workload runs in Oracle's public Frankfurt or Ashburn regions, it runs identically in the region on your raised floor.

Who buys it

The buyer profile is narrow and well defined. Across the sovereignty conversations we have had in 500+ OCI engagements, demand comes from five groups.

Strict sovereignty and in country mandates

Some jurisdictions require certain data classes to remain not just in country but under direct national or organizational control, in facilities the organization itself secures. National banks, sovereign wealth funds, and critical infrastructure operators in countries without a public OCI region are the canonical case. Where a public region in the right country exists, the calculus shifts, and options like the OCI EU Sovereign Cloud may satisfy the requirement at far lower commitment. Where it does not exist, Dedicated Region effectively creates one.

Defense and government

Defense ministries, intelligence adjacent agencies, and classified workloads often cannot use shared infrastructure at all, regardless of jurisdiction. A region inside a government secured facility, with physical access controlled by cleared personnel, answers requirements that no public region ever will. Several publicly announced Dedicated Region customers are exactly this profile: national governments standing up whole of government cloud platforms inside their own perimeters.

Banks and regulated financial institutions

Banks answering to national regulators with strong localization expectations have been among the earliest and largest adopters. The appeal is being able to give a supervisor a one sentence answer to the question of where the core banking estate runs: in our building, operated under our physical security, with the cloud operating model on top.

Telcos and latency sensitive estates

Operators running network functions, charging systems, or trading platforms care about single digit millisecond latency to specific physical locations: the mobile core, the factory floor, the exchange colocation hall. A full region placed next to the workload removes the speed of light from the argument. Manufacturers with large plants and exchanges with proximity requirements buy for the same reason.

Organizations with large existing data centers

Finally, there is a pragmatic buyer: the enterprise with substantial owned data center capacity, sunk investment in power and cooling, and a board reluctant to write that off. Dedicated Region lets them modernize to a cloud operating model without abandoning the facilities, which often changes the financial comparison materially.

What you provide, what Oracle provides

The division of labor is clean on paper and worth understanding precisely, because the customer side of it is a real program of work.

You provide the facility. Data center space meeting Oracle's specifications, power at the contracted capacity with appropriate redundancy, cooling, physical security, and network connectivity from the region to your corporate network and to the internet. You also provide the staff who control physical access and the compliance posture of the building itself. If your facility does not meet specification today, the uplift work, power upgrades, cooling, floor loading, security zones, sits on your critical path and your budget.

Oracle provides everything inside the cage. The racks, the servers, the network fabric, the storage, the software, the control plane, and the people who operate it all remotely. Oracle monitors the region around the clock, applies patches and updates on the same cadence as public regions, replaces failed hardware, and expands capacity within the contracted envelope. Your administrators manage tenancies, compartments, identity, and workloads exactly as they would in a public region; they do not manage infrastructure.

The seam between the two is the part to negotiate carefully: who escorts Oracle field engineers, how hardware arrives and leaves, what happens to failed disks containing your data, and how the region behaves if the connection to Oracle's management infrastructure is interrupted. These are solved problems with standard answers, but the answers belong in the contract, not in assumptions.

The commercial shape

Dedicated Region is sold as a multi year commitment, typically four years or more, with a sizeable annual minimum spend. You are not buying hardware; you are committing to consume a minimum amount of OCI services each year, drawn down at the same public list rates and Universal Credits mechanics as any public region. Consume more than the minimum and you pay for the overage at the same rates. Consume less and you pay the minimum anyway, which is the clause that makes capacity planning a board level topic.

The entry point has come down substantially. The original generation required a footprint and an annual minimum that put it firmly in the territory of the largest enterprises and governments. The current generation, introduced as Dedicated Region 25, shrinks the minimum physical footprint to a few racks and lowers the entry commitment accordingly, bringing the offering into reach for mid sized banks, single country telcos, and regional government bodies. Treat any specific number you have heard with caution: the commitment is set at contract time, varies with footprint, services, and geography, and is best characterized as a large multi year minimum that has come down with the smaller generation and must be confirmed with Oracle during negotiation. What you can rely on is the structure: minimum annual consumption, public region rates, and a term long enough for Oracle to recover the cost of building you a region.

Dedicated Region vs Exadata Cloud at Customer and Compute Cloud at Customer

Oracle sells three distinct ways to run its cloud on your premises, and conflating them is the most common error we see in early stage evaluations. Exadata Cloud at Customer brings the Exadata database service, and only that service, into your data center, managed by Oracle and connected to a public OCI region for its control plane. Compute Cloud at Customer does the same for a rack scale slice of compute, storage, and networking. Both are single service appliances with a cloud operating model. Dedicated Region is the whole region. The comparison below also includes the public region baseline, because that is the option every alternative must beat.

The decision rule that falls out of the table: if the constraint applies only to the database tier, Exadata Cloud at Customer solves it for a fraction of the commitment. If the constraint applies to the whole estate, or you need the full platform catalog inside your perimeter, Dedicated Region is the honest answer. And if your real ambition is to resell cloud services to others under your own brand, you are shopping for a different product entirely, Oracle Alloy, which we cover in our article on Oracle Alloy and becoming a cloud provider.

Realms, identity, and integration

A Dedicated Region joins an OCI realm, and the choice of realm is made at contract time. Most commercial customers join the standard commercial realm, which means their dedicated region appears alongside any public regions in the same tenancy, with native cross region networking, shared identity, and a single pane of administration. Government customers can attach to government realms, and isolated configurations exist for buyers who want no connectivity to any other region at all. The realm decision determines who you can peer with, where your identity domain lives, and how a hybrid estate behaves, so it belongs in the architecture conversation from day one, alongside the data classification and control mapping work we describe in our guide to architecting for data sovereignty.

For the common case, commercial realm membership, integration is pleasantly boring. Identity federation, compartments, tagging, budgets, and policy all extend to the dedicated region as they would to any new region subscription. Existing landing zone code typically needs region parameters and availability domain names updated and little else, which is one of the strongest practical arguments for the offering: your platform engineering investment carries over intact.

The operational model once it is running

Day two is where Dedicated Region either earns its commitment or does not. Oracle's remote operations team runs the region continuously: monitoring, patching, firmware, capacity expansion within the contract, and incident response on the infrastructure, all to the same SLAs published for public regions. Your teams keep everything they would own in any OCI region: tenancy security, identity, network design, workload operations, cost management, and backup and disaster recovery strategy. You also gain responsibilities no public region customer has: facility uptime, physical access management, escorting field engineers, and the network links into the region.

Plan the disaster recovery question early. A single dedicated region is a single region; if the workload class that justified it also demands regional DR, you need either a second dedicated region, a public region you are permitted to fail over to, or an in building strategy across availability domains with eyes open about what it does and does not protect against. None of these are wrong, but the choice changes the contract size. This is also the stage where an experienced partner pays for itself: our OCI implementation practice builds the landing zone, the connectivity, and the operating runbooks for dedicated regions the same way we do for public ones, with 24/7/365 monitoring available once the estate is live.

Risks and gotchas

Capacity planning is your problem again, at the contract level. Public cloud's deepest comfort is that someone else worries about capacity. With a dedicated region you commit to a minimum and a footprint years in advance. Undershoot and you pay for consumption that never happened. Outgrow the footprint and expansion is a hardware delivery and contract amendment, not an API call. Demand forecasting discipline, the skill many organizations were happy to retire, comes back.

Data center readiness drives the timeline. The interval between signature and a live region is measured in months, and facility readiness is usually the long pole: power upgrades, cooling, security buildout, and circuit provisioning routinely take longer than the region installation itself. Start the facility assessment before the commercial negotiation, not after.

Exit is a migration, not a cancellation. At the end of the term the hardware is Oracle's and the region can be decommissioned, but your workloads, data, and operating model need somewhere to go. An exit plan, even a rough one, belongs in the original business case: which public region, which realm, what data egress path, and what the runbook looks like. Regulators in financial services increasingly ask for exactly this document.

Catalog parity is real but not instantaneous. New OCI services and features reach dedicated regions on a rollout schedule. If a roadmap service matters to your plan, get its availability for dedicated regions in writing rather than assuming day one parity.

An eight step evaluation and readiness sequence

Before any serious conversation with Oracle, run this sequence. It exists to make sure the cheapest sufficient option wins, whatever that turns out to be.

Dedicated Region is the clearest expression of Oracle's distributed cloud strategy: rather than asking the customer to come to the cloud, it brings the entire cloud to the customer, at public region prices, in exchange for commitment and patience. For the defense ministry, the national bank, the telco core, and the enterprise whose data genuinely cannot leave the building, it solves a problem nothing else on the market solves as completely. For everyone else, the discipline is to prove the smaller options insufficient before signing the larger one. Either way, the architecture and the commitment number deserve independent scrutiny before the ink dries, because the decisions made at contract time are the ones that drive the typical 40% spend reduction we find, or fail to find, years later.