1. Choose the cloud that’s right for your workloads

Many companies want to move to the cloud, but it can seem too hard to get there. “They can’t just move their mainframes there,” Magouyrk said.

While tools to upgrade or rewrite business applications for cloud native environments are available, Oracle has made its public cloud “flexible enough to provide a truly virtualized environment, similar to what customers are used to on-premises,” he said. “We have a pervasive bare metal offering, virtualized Layer 2 networking, and block storage services that provide the same types of features as on-premises NAS and SAN devices, so when you want to move your applications to the cloud, you can.”

For hybrid cloud environments, Magouyrk said, many vendors just provide a massively scaled down subset of the technology that customers can deploy in their data centers. “Take AWS Outposts, you get five or six services that’ll run in your data center,” he said. “But customers want the benefits from all services that are available in the public cloud.”

Oracle’s approach with Dedicated Region Cloud@Customer is different, providing a complete public cloud operated by Oracle but run in each customer’s own data center. “We offer all of our 60 infrastructure services, as well as our full portfolio of cloud applications inside your data center,” Magouyrk said. “It’s a huge departure from how the rest of the industry has tried to approach the hybrid cloud.”

2. Let developers spend more time writing code

In the early days of cloud infrastructure, the foundational offerings spanned compute, storage, and virtualized networking services. But infrastructure technology has moved up the stack, with new offerings, such as managed database and application services.

“As the overall service portfolio has become more specialized, it just makes it so much easier for developers to spend more time writing code for the applications that matter to their businesses, and a lot less time configuring, deploying, managing, and maintaining them—all that undifferentiated heavy lifting,” Magouyrk told The Six Five Summit.

3. Move from custom to ‘pre-canned’ machine learning (ML)

Once the dominion of highly trained data scientists, machine learning tools and big data platforms are giving way to “cognitive services” and prebuilt data models, which any developer can integrate with a company’s own data sets and then customize and tune.

“This is a massive sea change that’s happening across the cloud infrastructure industry,” Magouyrk said. “It’ll be an accelerator to bring machine learning to the vast majority of applications.”