Mathias Rosenthal - Fotolia
Set your AWS cloud migration strategy before the big move
A move from on premises to the cloud doesn't happen with a click of a button. Set clear goals, and familiarize yourself with AWS' various native tools to make the most of your move.
AWS has several ways to move your applications and data to its cloud, but the transition can be painful if you don't have a proper cloud migration strategy.
IT managers don't simply copy over the data and hit the start button. They need to map out how their applications, data management, security infrastructure, existing integrations and configuration settings will work in the cloud in order for the process to go smoothly. It's also important to decide upfront whether to move everything to AWS or build hybrid cloud architecture for data and application mobility.
Every organization has different needs and objectives, and its cloud migration strategy will depend on how it wants its workloads to operate going forward, said Grant Kirkwood, CTO at Unitas Global, a Los Angeles-based hybrid cloud provider that works with AWS.
"The most important thing companies can do when thinking about migration is to clearly define and articulate the desired end state," Kirkwood said.
Pick an application migration strategy
First, choose an application migration strategy. It could be as simple as rehosting an application and copying its data to a comparable AWS offering. This is the quickest, simplest and cheapest approach, though it may not result in operational cost savings.
A slightly more complicated cloud migration strategy is to replatform an application. For example, you could move an existing managed database to a managed cloud database, such as Amazon Relational Database Service (RDS). You can retain the same architecture and shift management overhead to AWS, but it does require more configuration work.
The most difficult migration involves refactoring applications to take advantage of new programming models and services, such as AWS Lambda. This can increase scale and performance and lead to operational cost savings. But it's also the most costly and time-consuming and carries a bigger risk of failure due to the degree of change.
Inventory existing apps and track migration
Next, identify existing applications, configurations and interconnections. An enterprise may have thousands of workloads that run across multiple applications, many of which depend on other applications to function. AWS Application Discovery Service (ADS) identifies existing on-prem workloads and their underlying applications, collects configuration settings, assess usage and maps dependencies. This can help identify the total cost of ownership to run these servers on AWS and help organize a migration project.
Organizations should integrate the data AWS ADS collects into AWS Migration Hub to track the migration process. This provides a single view into an enterprise's migration process across multiple migration tools.
AWS Migration Hub helps ensure that all the subtasks in a migration are properly executed. For example, a migration might involve more steps than just the transfer of an application, including a database migration, the setup of virtual web servers and a shift from a bare-metal server or hypervisor to an EC2 instance. You might even need to update the load balancer to redirect traffic to the new instance.
Replicate the database
Organizations need to move their data first, unless they intend to connect migrated apps to an on-prem database. However, this hybrid approach will affect performance and incur higher data transit charges. AWS Database Management Service (DMS) captures a snapshot of data from a database without downtime, so the source database and dependent applications stay operational during the migration. This service works with most widely used commercial and open source databases.
AWS DMS can move data between similar databases, such as an on-prem Oracle Database to an Oracle Database hosted on Amazon RDS. You can also use it to transform and load data into a different database, such as a replatform move from Oracle to Amazon Aurora.
You can also replicate this same data into a consolidated data warehouse by streaming incoming data into Amazon Redshift or S3.
AWS DMS supports a wide variety of scenarios, including database consolidation, transfers to on-prem equipment and disaster recovery. It can also replicate an existing database for test and development. However, managers should mask sensitive data, such as credit card numbers in the U.S. or personally identifiable information in the EU, in the test environments.
Move the apps
IT teams can use AWS Server Migration Service (SMS) to help schedule, move and track incremental replicas of live server volumes. The service can automatically generate a library of Amazon Machine Images that match the characteristics and configurations of existing enterprise workloads captured by AWS ADS.
After you replicate the database to the cloud, you need to configure the identity and access management settings for AWS SMS. This enables AWS ADS to read the app from the existing server and then copy it to a comparable cloud instance. At the end of the process, you'll also need to reconfigure the domain name system records or load balancer to redirect to the new instance.
Organizations can also use VMware Cloud on AWS, a hybrid cloud service that can migrate vSphere-based environments between AWS and private infrastructure. An enterprise can use its existing VMware management console across cloud and private infrastructure and reduce the need to train operations support teams.
A recent AWS acquisition brought the CloudEndure Workload Mobility technology into the fold, too. This service automates more of the processes, including complexities around the migration of data, apps and configuration settings. Enterprises can use it to think about migrations at the workload level as a consolidated entity, which reduces the complexity.
Hybrid cloud and the big picture
This migration process gets more complicated when organizations want to migrate as part of a hybrid cloud strategy in which workloads share a common data source across AWS, other cloud providers and private infrastructure. This requires two key things: The data is in an open format, and it is accessible via the network.
Still, even a single-cloud migration can be a complex endeavor. Although all these tools seem simple on the surface, there are a lot of ways that a migration project can go wrong. To help identify and overcome these problems, enterprises might consider AWS Migration Acceleration Program (MAP). This is a professional service that helps organizations assess their readiness, do an initial plan and execute the actual migration. AWS MAP also works in collaboration with a wide variety of independent professional services firms that may have more experience with similar migration projects.
