Your Oracle estate has three futures. Choose one per workload.
Oracle now runs Exadata and Autonomous natively inside Google Cloud data centers. AlloyDB and BigQuery offer a path off Oracle entirely. The hard part was never the technology. It is knowing which of your databases belongs where, and in what order. That is where Dito starts.
Google Cloud regions targeted for Oracle AI Database@Google Cloud, announced at Next '26
Year Dito was founded. Google has been our only cloud ever since
Oracle database services available natively: Exadata, Autonomous AI, and Base
Unified bill, with usage counting toward your Google Cloud committed spend
The Oracle question stopped being binary.
For a decade the choice was stay on premises or commit to a rewrite. Neither answer survives contact with a real estate: one ERP instance nobody will touch, forty schemas of custom PL/SQL, an Exadata refresh due in eighteen months, and a support agreement no one wants to reopen. Applying a single strategy across all of it is what turns a migration into a multi-year detour.
Modernize everything
The team commits to PostgreSQL across the board, then meets the PL/SQL packages that run month-end close. Two years later, half the estate is still on Oracle and the operation now spans two platforms and two skill sets.
Move nothing
Deferring the decision is still a decision. It funds another hardware cycle, another support renewal, and another year of capacity planning against a roadmap set somewhere else.
Relocate it unchanged
The database lands in the cloud with the same operating model attached. Infrastructure line items move from capex to opex, the analytics and AI work stays out of reach, and the business case never arrives.
Most teams are working from an outdated picture of what is possible.
The options changed faster than the market's understanding of them. Before the strategy conversation is worth having, it helps to correct the assumption almost everyone brings to it.
“Oracle Database@Google Cloud is a fast network connection between Google Cloud and Oracle's cloud.”
Oracle Cloud Infrastructure database hardware, physically deployed inside Google Cloud data centers.
You provision it from the Google Cloud console. It connects to your Virtual Private Cloud over internal, low-latency network fabric. Applications on Google Kubernetes Engine or Compute Engine reach the database the way they reach anything else in the region. Cloud-to-cloud latency is not reduced here. It is removed, because there is no second cloud to cross.
Three landing zones. Every Oracle workload maps to one.
Dito assigns each database in your estate to a target before any migration work begins. The assignment is driven by what the workload actually depends on: Oracle-specific features, licensing exposure, application coupling, and where the data needs to be consumed next.
Run Oracle as Oracle
Oracle AI Database@Google Cloud
Oracle hosts, operates, and manages its database services inside Google Cloud data centers. Your teams keep the Oracle features they depend on and gain low-latency adjacency to Google Cloud compute, analytics, and AI.
- Exadata Database Service and Exascale infrastructure
- Autonomous AI Database, with cloning and Autonomous Data Guard
- Base Database Service on hardware-generation-agnostic x86 shapes
- Zero Data Loss Autonomous Recovery Service and Oracle Data Safe
- Oracle GoldenGate Service via Google Cloud Marketplace
- Terraform provider for repeatable, reviewed provisioning
RAC, Data Guard, or Exadata-specific capability is load bearing. The application is packaged and cannot be rewritten. The clock is set by a hardware refresh or a data center exit.
Keep Oracle, move it closer
Compute Engine, GCVE, Cross-Cloud Interconnect
Bring existing Oracle licenses to Google Compute Engine, or keep VMware-based estates intact on Google Cloud VMware Engine. Where OCI remains part of the picture, Cross-Cloud Interconnect links the two environments without cross-cloud data transfer charges.
- Existing licenses applied to Compute Engine instances
- Oracle E-Business Suite, PeopleSoft, and JD Edwards application tiers
- Google Cloud VMware Engine for estates that move as a unit
- Oracle Interconnect for Google Cloud between OCI and Google Cloud
- GKE and Compute Engine for the surrounding application layer
You need infrastructure-level control, the version or configuration is unusual, or the application tier has to move first with the database following on its own schedule.
Modernize off Oracle
AlloyDB, Cloud SQL, BigQuery, Spanner
Database Migration Service converts Oracle schemas, tables, and code objects into PostgreSQL syntax with Gemini-assisted conversion, then replicates continuously from the source until you are ready to cut over.
- AlloyDB and Cloud SQL for PostgreSQL as transactional targets
- Gemini-assisted schema and code conversion with issue explainability
- Sources include self-managed Oracle, RAC, and Bare Metal Solution
- Continuous change data capture keeps downtime to the cutover window
- BigQuery and Looker for the analytical estate
License cost is the binding constraint, the workload is a reasonable PostgreSQL fit, or the data is headed for BigQuery and the Oracle instance is really a staging area.
Scope note: Database Migration Service supports Oracle 12c through 21c as a source and does not support Oracle Autonomous Database as a migration source. Feature availability varies by region and service. Dito validates both during assessment.
Three native Oracle services. They are not interchangeable.
Exadata, Autonomous AI, and Base Database Service are all provisioned and integrated through the Google Cloud platform. Choosing between them is a design decision with cost and performance consequences that persist for years, and the common error is defaulting to the largest option because it feels safest.
Oracle Exadata Database Service@Google Cloud
Purpose-built, dedicated Exadata hardware provisioned inside Google Cloud facilities. This is the option for workloads where database performance is a business constraint rather than a technical preference.
You provision and orchestrate Exadata infrastructure directly in the Google Cloud console. The service links to your VPC through a local Oracle Database network running on internal, low-latency fabric.
Mission-critical systems that depend on multi-node RAC clustering and sub-millisecond response, including applications already running on Google Cloud VMware Engine or Compute Engine.
Dedicated Infrastructure
A physically isolated rack, minimum two database servers and three storage servers, delivering predictable performance equivalent to bare metal.
Exascale Infrastructure
A virtualized, shared elastic block storage tier that lowers the cost of entry while retaining Exadata's performance characteristics.
Oracle Autonomous AI Database@Google Cloud
Oracle's self-tuning, self-patching database running on Exadata infrastructure inside Google Cloud data centers, deployed without VMs, cluster software, or patch cycles for your team to own.
Serverless deployment. AI Vector Search is available without additional setup, so enterprise data can feed Vertex AI and Gemini models directly. GoldenGate streams the same data into BigQuery for analytics.
Teams that want Oracle's capability without Oracle's operational overhead, and estates where the database is about to become an input to AI and analytics work rather than only a system of record.
Oracle Base Database Service@Google Cloud
A VM-based Oracle database running on general-purpose OCI hardware co-located in Google Cloud data centers, on x86 shapes designed to be hardware-generation agnostic.
Standardized compute profiles and scalable block storage, connected to your Google Cloud environment through internal VPC peering, running alongside your existing Google Cloud VMs.
Development and test environments, smaller applications, and systems outside the mission-critical tier, with full compatibility for standard Oracle tooling including RMAN and Data Guard.
One invoice, against your committed spend
These services are purchased through Google Cloud Marketplace. Usage counts toward your Google Cloud committed spend and arrives on a single unified bill.
The access controls you already operate
Database access maps to standard Google Cloud Identity and Access Management, so your security posture stays centralized instead of forking into a second model.
The tooling your team already uses
Database metrics surface in Google Cloud Observability, and resources are managed through standard Google Cloud APIs and command line tools.
Dito's assessment names the service, the infrastructure option, and the sizing for each workload, then models what each choice costs against your committed spend before anything is provisioned.
What changed for Oracle on Google Cloud this year.
The 2026 releases moved this partnership past hosting. Oracle data is now addressable by Google Cloud governance, analytics, and agents without leaving the database. These are the capabilities we design around today.
Regional expansion to 20 regions
Announced at Google Cloud Next '26, extending data residency and latency options for regulated and multinational estates.
Oracle GoldenGate Service on Google Cloud
Low-impact replication for migrating on-premises Oracle databases, plus managed near real-time movement from Oracle into BigQuery.
Knowledge Catalog integration
Autonomous and Exadata workloads governed alongside Cloud Storage and BigQuery under one catalog rather than a separate Oracle silo.
Database Center fleet monitoring
Compute, storage, and network health across your Exadata and Autonomous inventory, queryable through a natural language interface.
Managed MCP servers
PreviewGemini and other MCP-compliant models interact with Oracle data through a managed endpoint, with no server infrastructure to host, secure, or scale yourself.
Oracle AI Database Agent
PreviewBusiness users query Oracle data in natural language. Identity propagates from the interface to the database, so results respect existing row and column level privileges.
Sources: Google Cloud, "What's new for Google Cloud databases at Next '26" and cloud.google.com/solutions/oracle; Oracle, "Oracle Expands Powerful AI Capabilities in Oracle AI Database@Google Cloud," April 2026. Preview features are subject to change and are not covered by the same service terms as generally available services.
Running Oracle on Bare Metal Solution? Plan the exit while you still control the sequence.
Google Cloud has moved Bare Metal Solution to a specialized, allowlist-only availability model and is directing existing customers toward the Oracle and Google Cloud partnership offerings. This is not an emergency. It is a window. Estates that map their exit now get to sequence it around business cycles instead of around a contract date.
Recommended next steps:
- Inventory every database, version, edition, and Oracle option currently in use
- Assign a target landing zone and a migration method per database
- Model licensing implications before committing to a target
- Sequence cutovers around fiscal close, peak season, and release freezes
- Validate rollback at each step, not just at the end
Oracle depth and Google Cloud focus, in the same practice.
Most partners have one or the other. Firms with real Oracle bench strength treat Google Cloud as a secondary platform. Google Cloud specialists tend to route the Oracle work elsewhere. Dito holds both, which is what the landing zone decision actually requires.
Oracle expertise held in house
Our practice is staffed by people who have carried the pager for Oracle in production: RAC, Data Guard, RMAN, GoldenGate, ASM, and the PL/SQL that never made it into anyone's documentation. The assessment work and the hard parts of the migration are not subcontracted.
Google Cloud is our only cloud
Dito has been a dedicated Google partner since 2007, working as a Co-Sell and Services partner across Google Cloud infrastructure, data, AI, and Google SecOps. That focus is why we can tell you what is generally available, what is in preview, and what is still on a roadmap.
Secure by design, not secured afterward
IAM boundaries, VPC Service Controls, customer-managed encryption keys, and Oracle Data Safe posture are designed into the target state before the first workload moves. Our Google SecOps practice reviews the result rather than inheriting it.
The DBA team decides whether this works
A fifteen-year Oracle team can make a migration succeed or quietly stall it. We plan for that explicitly: shared design sessions, hands-on enablement on the target platform, and runbooks written for the people who will actually be on call.
Assessment first. Nothing moves before the map exists.
Estate assessment
Full inventory of instances, versions, editions, options in use, custom code, integrations, and dependencies. Typically two to three weeks depending on estate size.
Landing zone design
Target architecture, network and identity design, and licensing analysis. One representative workload is migrated and validated before the plan is finalized.
Migration at scale
GoldenGate or Database Migration Service replication, parallel run, data validation, and a rehearsed cutover with a tested rollback path.
Run or hand over
Dito operates the estate as a managed service, or trains your team and transfers runbooks, monitoring, and escalation paths. Your call, and it can change later.
What the assessment leaves behind
What technical teams ask first.
Can we keep our existing Oracle licenses?
Is this a link to Oracle's cloud, or is Oracle really running inside Google Cloud?
Do RAC, Data Guard, and Exadata features still work?
How long does an Oracle to PostgreSQL migration take?
What about E-Business Suite, PeopleSoft, and JD Edwards?
Is Bare Metal Solution still an option for new deployments?
Can Oracle data reach BigQuery without a full migration?
Start with the map, not the migration.
A Dito Oracle estate assessment gives you a landing zone recommendation for every database, the licensing implications of each option, and a sequenced roadmap you can take to your board. Whether you migrate with us afterward is a separate decision.
Dito is a Google Cloud Premier Partner. Oracle, Exadata, GoldenGate, and Autonomous Database are trademarks of Oracle Corporation. Google Cloud, BigQuery, AlloyDB, and Gemini are trademarks of Google LLC.
