Enabling Oracle hybrid multicloud solutions with Digital Realty

February 7, 2024 | 4 minute read
Mathew (Matt) Sebastian
Director, Oracle Solution Center
Text Size 100%:

This blog was co-authored with Rutger ter Hoeven, Alliances Director, Digital Realty.

Industry trends driving cloud-adjacent colocation solutions

According to Precedence Research, the global cloud computing market size is valued at over $500 billion in 2023, from $446.51 billion in 2022, and is expected to surpass $1 trillion by 2028 and $1.6 trillion by 2030. While this explosive growth stems from the growing number of global public cloud regions, many organizations prefer to opt for hybrid multicloud architectures. In fact, many customers are deploying split-stack architectures with data on private infrastructure interconnected to applications in one or multiple cloud platforms.

A hybrid cloud approach helps bridge the traditional on-premises data center and the public cloud, bringing public cloud benefits to on-premises applications. Some industries and countries have data residency and data sovereignty regulations mandating that certain data remain on-premises, while some countries aren’t yet supported by in-country public cloud data centers, warranting a hybrid-cloud approach.

Even without a regulatory requirement, multiple, tightly coupled on-premises systems aren’t uncommon. For example, financial institutions often settle transactions through a mainframe connected to front-office applications and a shared authentication server. Moving all those systems into a public cloud isn’t possible, and moving only some of them could severely impact performance. Keeping these connected systems on-premises is best practice, at least until you can move them all to the public cloud.

A popular solution to the data gravity challenge of keeping sensitive data in private infrastructure, with applications residing in one or multiple cloud platforms, is cloud-adjacent colocation. When organizations decide to work with infrastructure adjacent to the cloud, the possibilities of integrating with other products and services are countless. For example, an organization might choose to host their databases on Exadata Cloud@Customer within a private cage in a colocation data center and interconnect with Amazon Web Services (AWS) or Azure through low-latency connections.

Split-stack and full-stack architectures

Full stack architectures involve running the entire database and applications on the Oracle Cloud Infrastructure (OCI) stack. You can further extend the concept to a stack-stack architecture that involves interoperability between the OCI stack and another cloud service provider (CSP)’s stack running its own set of database and applications. Data integration between applications is made possible without rearchitecture. Latency requirements depend on whether the data integration pipeline involves batch processing or real-time online streaming. Throughput requirements depend on data transfer volumes.

On the other hand, split-stack architectures, involve running applications on one cloud while the database runs on another cloud. An example includes running an Oracle Database on OCI connecting to a Power BI analytics application running on Azure Cloud. By running application components in its most optimal environment and running Oracle databases on OCI, you can achieve superior performance and cloud economics. This architecture typically requires low-latency and high-throughput network connectivity to achieve optimal performance between the multiple cloud provider’s stacks.

Flexible Architectures with Multicloud
Figure-1: Flexible architectures available with multicloud


Digital Realty ServiceFabric on PlatformDIGITAL

ServiceFabric, Digital Realty's service orchestration platform, seamlessly interconnects workflow participants, applications, clouds, and ecosystems and is deployed in the PlatformDIGITAL global data center platform. Two important and common uses for ServiceFabric are multicloud and port-to-port connectivity. ServiceFabric enables fast deployment of connectivity to private and public clouds using one or multiple ports to connect to any third-party ports, vRouters, or CSPs anywhere at any time through a global platform.

Why Digital Realty and Oracle?

The Digital Realty hybrid multicloud solution for OCI optimizes applications and database workloads for complex hybrid and multicloud environments. This solution involves on-premises infrastructure placed in Digital Realty's colocation centers with connectivity to OCI and other multicloud service providers. Digital Realty's ServiceFabric helps customers address the following issues:

  • Seamless low-latency, high-performance hybrid multicloud connectivity
  • Data colocation and proximity
  • Data sovereignty
  • Hybrid multicloud split-stack architectures

Using ServiceFabric in a split-stack hybrid multicloud architecture

The split-stack reference architecture shown in Figure 1 depicts Exadata Database Service running on Exadata Cloud@Customer connected to OCI using ServiceFabric and OCI FastConnect. It backs up all the data to an object store using a service gateway on the public cloud or OCI. ServiceFabric enables the Azure interconnection to a horizontal application that runs Azure Power BI analytics. This scenario is popular with customers who opt to have their physical database infrastructure in a cloud-adjacent colocation cage powered by Digital Realty, with the split-stack application running on the cloud.

Split-Stack Architecture with Exadata Database Service running on Exadata Cloud@Customer in a Digital Realty Cloud-Adjacent center connected to OCI and Azure stack via Service Fabric and OCI Fastconnect and Azure ExpressRoute
Figure-2: Split-stack architecture with Exadata Database Service running on Exadata Cloud@Customer in a Digital Realty colocation center connected to OCI and Azure stack via ServiceFabric - OCI Fastconnect and Azure ExpressRoute


Digital Realty’s cloud-adjacent colocation enables the following capabilities:

  • Data proximity to OCI and other leading cloud platforms
  • Running data analytics at the source
  • Modernizing data estate operations, enabling real-time data analytics, faster delivery of data reports, and more informed and timely business decision-making
  • Submillisecond split-stack performance and high throughput
  • Reduction of data transfer rates from several hours to few minutes for complex hybrid multicloud data replication operations


Want to know more?

To see these solutions in action, contact the Oracle Solution Center through your Oracle account representative and ask how to engage the Oracle Solution Center in validating any of these solutions before you invest.

Connect with the authors:

Matt Sebastian: https://www.linkedin.com/in/mathewsebastian/

Rutger ter Hoeven: https://www.linkedin.com/in/rutger-ter-hoeven-ab43931/ 



Mathew (Matt) Sebastian

Director, Oracle Solution Center

Matt leads the Partner /System Integrator (SI) Solutions within Oracle's RevOps Solution Center. He is responsible for global solutions and consulting for North America Sales, SE/SC orgs, Partners (GSI/RSI, VAD/VARs, ISVs and NSPs) focused on delivery, growth and adoption of OCI, Hybrid/Multi-Cloud, Cloud Adjacent and Engineered systems. 

Previous Post

Customer security best practices for Fusion application deployment

David B. Cross | 4 min read

Next Post

Oracle Cloud Infrastructure Logging Analytics best practices series

Royce Fu | 3 min read