Oracle Corp. and Amazon Web Services Inc. seem to be putting their longstanding rivalry aside in favor of a more pragmatic relationship that acknowledges the reality of today’s multicloud technology environments.

The two technology giants today revealed a plan to establish private, high-speed connectivity between Oracle Cloud Infrastructure and the AWS cloud. Joint customers will be able to move data and run applications across them with reduced latency. By establishing a direct connection between Oracle Interconnect and AWS Interconnect-multicloud, the two companies can offer companies a fully managed, enterprise-grade pipeline that effectively transforms their respective cloud platforms into a unified computing environment.

The move reflects the enterprise shift toward “multicloud” infrastructures that make use of dozens of individual cloud platforms and software services. Multicloud is not new, of course, but until recently organizations have struggled to properly integrate different clouds with one another.

Historically, they’ve been forced to build their own connections, and have relied on a messy patchwork of third-party network providers, manual configurations and high-priced physical infrastructure. This has resulted in some major “data gravity” issues, where data effectively becomes trapped in one cloud platform because it’s far too expensive or complex to move it where it’s needed.

These problems have only worsened with the rise of artificial intelligence workloads. Modern AI applications usually run on a split-stack architecture – for example, a company might use a high-performance database on Oracle’s OCI together with AWS SageMaker for application logic and model training. It’s not easy to change things once they’re up and running, and without a high-performance bridge between the two clouds, such workloads can suffer from extreme latency, limiting their usefulness.

This is the challenge that Oracle and AWS now want to solve with today’s announcement, and it’s a fairly surprising one given that the two companies have long been extremely critical of one another. However, it seems that both stand to benefit by becoming “frenemies” and enabling data to flow between their respective clouds as if it were a single, integrated platform.

To that end, Oracle is integrating its native interconnect service with the AWS Interconnect-multicloud specification to create a secure and private connection that bypasses the public internet completely. Doing so will enable the companies to support split-stack deployments, so organizations might run an application on OCI and store its data on AWS, or vice versa.

Alternatively, they now have a way to move all of their data on one cloud to another, without the usual complexity that comes with such migrations. Notably, Oracle said the new connection is “managed” and “native,” which means there’s no need for customers to set up manual routing or design complex data replication strategies. By plugging into AWS’ specification, Oracle is essentially standardizing how its cloud talks to its rival’s platform.

Rob Strechay, principal analyst of theCUBE Research and Smuget Consulting, told SiliconANGLE that today’s move underscores the belief that multicloud is no longer just a strategy, but a reality for every enterprise. “By removing the networking complexity between two of the largest domains, Oracle and AWS are making cross-cloud resilience and AI architectures far more achievable for enterprises,” he said.

The enhanced cloud connectivity will likely accelerate the adoption of multicloud environments even more. Strechay said the complexity of networking has been one of the biggest challenges stopping more enterprises from embracing split-stack environments. Enterprises need a solution, he added, because the future of AI is almost certainly not going to be in single-cloud deployments.

“The future of AI is where your data lives in one place, your models run in another, and the network doesn’t get in the way,” Strechay said. “This also enables organizations to easily cross-cloud disaster recovery. While it has always made sense for databases on paper, Oracle is making it operationally achievable without building a networking science project.”

The announcement is not without precedent. Another rival, Google Cloud, has long been open to the realities of multicloud, launching its Cross-Cloud Interconnect service as far back as May 2023, before following with a dedicated private connection to AWS last December. For its part, Oracle has already established interconnects with Google Cloud and Microsoft Corp.’s Azure.

Nathan Thomas, Oracle’s senior vice president of product management, said today’s announcement builds on an earlier collaboration between the two cloud giants that led to the launch of Oracle Database@AWS, which made it possible to run Oracle database workloads on the AWS cloud more easily. “This will help our mutual customers modernize their applications, unify their data, and unlock new generative AI opportunities,” he promised.

The high-performance link is meant for enterprise-scale workloads and will leverage Oracle’s network of 26 interconnected partner cloud regions. It’s expected to go live later this year, with initial availability in the AWS US East (N. Virginia) region.

Image: SiliconANGLE/Gemini