If your company is looking to transform systems that have an overwhelming quantity, diversity, and/or complexity, you fall under the umbrella of “large scale migrations.” These situations often involve multiple business units, and there are very demanding operational and security requirements in production workloads.

Randall Barnes, a Principal Cloud Architect for Oracle, has helped countless companies transform their systems and processes over the years—saving cash and increasing capacities. This experience has allowed him to identify themes and patterns within large scale Oracle Cloud Infrastructure migrations. During his presentation at COLLABORATE 19, Randall confessed his mistakes and shared his victories to make sure others have as easy of a transition as possible.

Common Struggles and Solutions for Large Scale Migrations

Common struggles that occur during large-scale migrations include:

• Trials and tribulations of consensus building and inter-team coordination
• Unraveling legacy system dependencies
• Data transfer and sync requirements

To address these struggles, Randall shares four main strategies:

1. Rehost
2. Rearchitect
3. Replatform
4. Retire

Rehosting is a “lift and shift” technique. This involves both virtual and physical systems. Users pack up current systems, transfer environments, and relaunch or configure changes to make sure systems are in sync. The major advantage of rehosting is encapsulation. The disadvantages include a limit on Cloud support for certain platforms, potential performance degradation in emulated mode, and licensing hindrances.

Rearchitecting to OCI Cloud Native can look like taking a static number of application systems and refactoring them to take advantage of automatically scaling nodes or scaling virtual CPUs into a larger or smaller instance with some level of coding.

Replatforming with a PaaS or SaaS adoption is a popular approach. Major perks include availability for help from the PaaS or SaaS team, but potential problems may show up if specific features or versions necessary for your functions are not included.

Retiring current systems is another strategy for migration. Often, customers begin by creating a spreadsheet of systems that they know they are using. During the assessment phase, we find that a lot of systems are orphaned, or they are only used once or twice a year. Ultimately, 25 percent of those systems are retired on the spot.

Steps for Oracle Cloud Infrastructure Migration

Most organizations’ migration will involve a combination of strategies. When your team is ready to implement changes, you should follow this six-phase approach:

1. Assessment
2. Define target architecture, migration strategy, and cost estimates
3. Develop RACI and Execution Schedule
4. Deploy foundation infrastructure and shared services
5. Data transfer and hydration
6. Execute scheduled waves

Large-scale migrations are complicated, but when you follow best practices and learn from the mistakes of others, you will be on track for a successful and streamlined transition.