Inside Oracle’s cloud adoption journey

Migration strategy

With more than 1,000 applications under her charge at Oracle, Evans had her work cut out to rationalise which applications were still needed. Her team found that in some cases, three different applications had been created by various business units.

“We created a central repository of our assets and once we got a good sense of what was being used and how it was being used – some applications weren’t even used – we could get a level of transparency while we did the migration,” Evans said.

Her team also realised that a significant amount of infrastructure being provisioned was underutilised. So, by migrating its infrastructure to OCI in a ‘lift-and-improve’ approach, Oracle was able to reduce waste and achieve infrastructure efficiency that resulted in cost savings.

Evans said although a lot of lift-and-improve work was done, new containerised applications were also developed and deployed using OCI’s cloud-native tools. In some cases, applications such as a monolithic subscription management platform had to be broken down and rewritten as microservices to achieve the scale Oracle needed to make sure its customers were truly paying for the services they used. “That was a big rewrite, but we also optimised it for the cloud and continued to iterate from there,” Evans said.

At the same time, the use of managed services helped with tasks such as database tuning. “With autonomous database capabilities, we can do things like tuning, scaling, and repairing if you have an incident or service impacting your database, without having our database admins do those tasks anymore,” Evans said.

Underpinning Oracle’s cloud migration strategy was a reference architecture that outlined application tenancies, how applications interacted with each other, and security controls, alleviating the need to do bespoke reviews with application, security, operations and architecture teams.

To keep cloud costs in check, Evans said Oracle uses the Oracle Cloud Advisor to understand which workloads are being run, and what’s running in terms of compute, memory, storage and databases, so it can make decisions to manage costs.

For example, monitoring, logging and telemetry, while necessary to keep applications and services humming along, take up a lot of resources. “But at the same time, you don’t need to have all of that, and so we’ve compressed a lot of the tracking to only track things that you absolutely need to.”

Besides being Oracle’s CIO, Evans also oversees platform services for OCI, with her experience deploying OCI services translating into product improvements.

For example, while using Oracle’s generative AI (GenAI) services including pre-trained models to create an AI service desk that could better address the surge in password-reset requests at the start of each year, Evans’ team encountered performance issues which they worked with Oracle’s GenAI product team to fix.

Oracle has also developed fleet management tools to gain better visibility over cloud resources used across its business units, enabling it to provision and manage workloads in an automated and centralised way. “We’ve created the tools for ourselves, but now we're in the process of making them available as a service to our customers,” she said.