Oracle has expanded its alliance with Amazon Web Services (AWS) to provide a set of high-speed interconnects between their respective cloud services.

Muneer Mirza, vice president of multicloud for the Oracle Cloud Infrastructure (OCI) service, said the level of networking integration enabled is similar to the integration that Oracle has also made with Google and Microsoft to facilitate integration of applications spanning hybrid cloud computing environments.

The AWS Interconnect is based on an open specification defined by AWS. In total, however, there is now OCI connectivity spanning 26 interconnected partner cloud regions.

In each case, the integration can be achieved by the same IT team that provisioned the cloud service rather than requiring networking expertise, said Mirza.

The overall goal is to make it simpler for IT teams to deploy workloads on different clouds based on the requirements of specific workloads, he added.

While AWS, Microsoft and Google still dominate the cloud computing space, Oracle has been making some gains. As organizations invest more in artificial intelligence (AI), more of them are considering OCI services that have long been optimized for managing massive amounts of data, noted Mirza.

Tom Hollingsworth, an industry analyst with the Futurum Group, said the alliance creates an opportunity for Oracle to capture organizations that want to explore AI but have most of their data located on AWS. The fast, private interconnect provides a way to keep data safe as it is transferred into Oracle’s cloud for AI processing, he added.

In general, IT organizations are making more nuanced decisions when it comes to deploying workloads. Not all cloud computing environments run all types of workloads equally well. In addition, there is still a large amount of data stored in an on-premises IT environment that may need to be exposed to compute engines running in the cloud. The end result is a greater need for high-speed interconnects to minimize the impact network latency might have on overall application performance.

Ultimately, most organizations are going to find themselves deploying workloads on at least two cloud computing environments if for no other reason than because a merger or acquisition forced the issue. While the goal might be to deploy as many workloads as possible on a preferred cloud, the fact remains there will almost always be a business unit or department that has, whether IT likes it or not, elected for one reason or another to deploy workloads on a different cloud service.

The one thing that is clear is that as networking is managed at higher levels of abstraction, the responsibility for managing networks is being subsumed into a larger cloud operations team that is relying on a cloud service provider to manage the underlying infrastructure. If and when there is an issue, of course, the assumption is the cloud service provider has now assumed responsibility for rerouting traffic in a way that ensures service level agreements (SLAs) are still being met regardless of what issues might be impacting any individual router or network switch.