Oracle introduced a hybrid-cloud service designed to place Oracle’s managed database experience inside customer-selected locations.
Base Database Cloud@Customer extends Oracle's current Base Database Service via Oracle Cloud Infrastructure (OCI) and Oracle’s multicloud arrangements into customer data centers, branch locations, factories, and remote sites.
The service works in contrast to Oracle's Exadata Cloud@Customer large enterprise offering by targeting mid-sized, business-critical workloads in which data and applications must be kept at customer locations due to data residency and regulatory requirements or low-latency connections to existing systems.
With it, customers run Oracle AI Database, Oracle Database 19c, and associated applications while Oracle manages the underlying infrastructure. Data remains at the customer’s premises and behind its firewall, but provisioning, scaling, patching, backup, and other life cycle operations are performed through an Oracle cloud control plane.
The hyperscaler's new service runs on Oracle Data Infrastructure Cloud@Customer X11, an eight-rack-unit (8RU) system containing two X11 compute servers and shared all-flash storage. Each server provides 60 usable x86 processor cores and 660 gigabytes of usable memory. Storage starts at 11.6 terabytes and can expand online to 47.2 terabytes, while databases can be deployed across both servers for resilience during maintenance or failures.
Applications can run in virtual machines (VMs) on the same infrastructure, reducing latency and avoiding separate server deployments. Disaster recovery and protection can also be added through Oracle Data Guard and Zero Data Loss Recovery Appliance, including ransomware-resilient recovery.
On the AI side, customers can run Oracle AI Database 26ai in Standard or Enterprise Edition. Oracle touted tools such as AI Vector Search, which creates vector embeddings, searches structured, and unstructured data semantically and supports retrieval-augmented generation (RAG); Private Agent Factory, a containerized no-code tool for building, deploying, and managing data-driven AI agents and workflows; and Private AI Services Container, which runs private AI models for bulk embedding, content generation, and inference using open-weight large language models (LLMs).
Application VMs can host AI agents beside the database, creating an end-to-end agentic AI environment behind the customer’s firewall, isolated from third-party AI providers and designed to meet security and data-residency requirements.
Base Database Cloud@Customer also offers online compute scaling and pay-as-you-go database pricing, allowing customers to reduce capacity during low usage periods, helping to cut down on resource-intensive AI inferencing.