AWS Migration Considerations Series: The Roadmap

 

We have covered how a workload is migrated to the Cloud and the actions we will have to start taking responsibility for as soon as a workload is live. Let's start planning out the activities we will do to get there.

For our AWS clients, we follow the Migration Acceleration Program approach, which defines three phases and a set of activities.

AWS Migration Acceleration Program Key Phases

These phases take time but increase organizational awareness of responsibilities and skill gaps. This awareness ensures that migrations do not happen as shadow IT and misalign with other initiatives.

Prioritizations and Waves

There is no magic “migrate everything and optimize it” button. Implementing a workload requires an engineer to understand the application, its dependencies, integrations, data models, users, and more.

Some of the items we review in workloads are:

• Number of inbound integrations and dependencies from other systems
• The frequency of these integrations inbound, protocols, and data volumes
• Number of outbound integrations and dependencies
• Frequency of these integrations outbound, protocols, and data volumes
• Number of users publicly (i.e., over the Internet, if any)
• Number of users internally
• Size of any data file store to migrate
• Size and type (Engine) of any database to migrate
• Licensing requirements.

Cloud Migration Optimization and Modernization

We look at things we can optimize and modernize at the same time, checking commercial-off-the-shelf (COTS) vendor support and corporate standards:

• Current app server operating system version and target operating system version
• Current database engine and version, and target database engine version
• Current installed application revision and latest patch revision
• Current service endpoint and suitability for placing behind a managed load balancer with a TLS certificate from the Amazon Certificate Manager
• Suitability for templating
• Service backup and restore options once migrated to the cloud.

A few initial workloads are generally low-priority, lightly integrated, or standalone systems. Some can run active-active from the cloud and on-premise, whereas others may have to run from a single location at any time.

Akkodis has been an AWS Consulting Partner since 2013. Learn more about our AWS Practice and services. 

By James Bromberger, VP Cloud Computing, Akkodis Australia