Migrate

In the migrate phase, we ensure the migration proceeds as planned, monitor the migration process, and have a plan in place to rollback in case any issue encountered during the migration. During migration, you can scale your resources corresponding to the volume of data to be migrated. Furthermore, you can adopt best practices that can reduce interim resource consumption during your migration.

MIG-SUS-07: Does your on-premises to AWS data migration strategy consider sustainability?

Data makes up the large portion of the scope of many workload migrations. Identifying and optimizing the data storage with latest technologies helps improve the power efficiency and reduce carbon footprint.

MIG-SUS-BP-7.1: Implement data management practices

This BP applies to the following best practice areas: Data

Data management is a continuous process and should be implemented during and after the migration. With the latest storage technologies, it provides the opportunity to configure and provision sufficient storage without compromising the business needs.

Implementation guidance

Suggestion 7.1.1: Avoid over-provisioning for storage system to influence your environmental impact.

• Perform application discovery to identify data characteristics and access patterns that can be supported by storage technology.

• You can use shared file systems or storage that allows for sharing data to one or more consumers without having to copy the data. For example, you can have a shared drive to store common files instead of copying those common files to each VM.

• After migrating the workload, from time to time, analyze data access and date movement to identify opportunities to become more efficient. When opportunities are found, change the lifecycle by moving to other storage classes or deleting unneeded data.

• Use technologies that support data access and storage patterns. For example, migrating data to other object storage types eliminates provisioning the excess capacity from fixed volume sizes on block storage. For more detail, see SUS04-BP02 Use technologies that support data access and storage patterns.

Suggestion 7.1.2: As part of your per-migration planning evaluate your current recovery time objective (RTO) and recovery point objective (RPO).

• Design your backup strategy based on your actual business requirements. Avoid backing up non-critical data that has no business value, and detach volumes from clients that are not used before considering to migrate those workloads. For more detail, see SUS04-BP08 Back up data only when difficult to recreate.

• Use an automated solution or managed service to back up business-critical data. AWS Backup is a fully-managed service that makes it easy to centralize and automate data protection across AWS services, in the cloud, and on-premises. Next to other capabilities, AWS Backup helps you become more sustainable. For example, you can use Backup to set an expiration on your manual snapshots.

• Set automated lifecycle policies to enforce lifecycle rules for the migrated data. For more detail, see SUS04-BP03 Use policies to manage the lifecycle of your datasets.

• If you are setting up disaster recovery for your migrating workload, evaluate your RTO and RPO, and see if you could meet the requirement using the backup data instead of replicating the entire data to the recovery site. For more detail, see AWS Elastic Disaster Recovery.

Suggestion 7.1.3: Choose the right migration tool, and scale your resources corresponding to the volume of data to be migrated.

• AWS provides migration services like AWS Database Migration Service and AWS Application Migration Service. You may be able to scale down the replication instance type selected if the amount and velocity of the ongoing data is much smaller than the amount of historical data.

• Another alternative is to use a serverless migration tool like AWS DMS Serverless.

• Here are some other options to choose from to migrate your storage with their key characteristics.

Migrate your storage	Key Characteristic
AWS DataSync	Simplify, automate, and accelerate data movement to and from AWS Storage, as well as between AWS Storage. Easily manage data movement workloads with bandwidth throttling, migration scheduling, task filtering, and task reporting with a fully managed service that seamlessly scales as data loads increase.
AWS Transfer Family	Simply and seamlessly move your files to Amazon S3 and Amazon Elastic File System (Amazon EFS) using SFTP, FTPS and FTP protocol. Store information in Amazon S3 or Amazon EFS, manage workflows, and initiate automated, event-driven tasks with a fully-managed, low-code service. Quickly scale your business-to-business (B2B) file transfers for each line-of-business user.
AWS Snow Family	Collect and process data at the edge, and migrate data into and out of AWS through physical devices and capacity points. Device options range to optimize for space- or weight-constrained environments, portability, and flexible networking options.

For more detail, see the following:

• Data lifecycle management

• Amazon S3 Intelligent-Tiering

• I/O characteristics and monitoring

• Optimizing your AWS Infrastructure for Sustainability, Part III: Networking

• Top 10 Data Migration Best Practices

• AWS Summit SF 2022 - Optimizing your AWS infrastructure for sustainability

• Amazon EBS and Snapshot Optimization Strategies for Better Performance and Cost Savings

MIG-SUS-08: Are you adopting practices that can reduce interim resource consumption during the migration?

During a migration, your consumption of resources may increase due to the provisioning of resources in both the source and target environments. The increase in consumption is often referred to as a double bubble. In addition, your consumption may also increase due to provisioning of migration resources, such as the networking between your source and target environments, SFTP servers, AWS Application Migration Service (MGN), or AWS Database Migration Service (DMS).

Typical migration process flow

You can reduce the resource consumption during the migration either by reducing the resources deployed or by reducing the duration of their deployment.

Additional resource consumption equation

MIG-SUS-BP-8.1: Adopt methods that can reduce interim resource consumption during the migration

This BP applies to the following best practice areas: Process and culture

Implementation guidance

Suggestion 8.1.1: Reduce interim resources created in the target environment.

• Reconsider the migration of development and other non-production environments, as these can be rebuilt when required in AWS. If you decide on migrating your non-production environments, revisit the portions of the environment that need to be migrated. For example, you may choose to migrate only some of the databases in a database server.

• If you decide to migrate your build environment, increase the utilization of these environments.

• Use managed device farms to test new features on a representative set of hardware.

• During the migration, consider the impact on sustainability of day-to-day operations. For example, consider avoiding frequent backups of the target environment and setting up HA or DR during the migration. You can also reduce the retention period of logs or backup snapshots taken during the migration.

Suggestion 8.1.2: Reduce interim resources used in the migration process.

• During a migration, you typically have to migrate historical and on-going data. Historical data refers to the data that was created prior to the start of the migration. On-going data refers to the new data that is generated in the source environment at the time of the migration until the cutover. The resource needs for the migration of historical data may differ from that of the on-going data. Choose the right migration process and tool, and also scale your resources corresponding to the data to be migrated. For example, in the case of AWS DMS and AWS MGN, you may be able to scale down the replication instance type selected if the volume and velocity of the ongoing data is significantly less to volume of historical data. Another alternative is to use a serverless migration tool like AWS DMS Serverless that can automatically scale based on the volume of data being migrated.

• Share your migration resources if possible. Some migration tools let you share migration resources. An example of this is AWS MGN, which automatically shares the replication instance with multiple source servers being migrated.

• For migration resources that cannot be scaled easily, such as the networking resources between your data center and AWS Cloud, consider flattening the demand curve using buffering and throttling to reduce the required provisioned capacity for the workload. For example, you can throttle your network in AWS MGN when migrating your servers to AWS.

• Review the need to include migration resources in your day-to-day operations. For example, avoid including AWS MGN replication servers in your backup strategy. If you are capturing logs for migration resources, you can consider reducing the retention period for these logs.

Suggestion 8.1.3: Reduce duration of deployment for the interim resources created during the migration.

• Consider selecting a partner who has the technical expertise and experience migrating to AWS

• Create a cross-functional cloud-enablement team to implement the governance, best practices, training, and architecture needed for cloud adoption. The team will define tools, processes, and architectures that establish the organizations cloud operating model. In addition, it will coordinate with stakeholders across different units such as infrastructure, security, applications, and business to alleviate obstacles in a migration.

• Explore tooling that can facilitate your migration and can automate and expedite aspects of the migration such as discovery, project management, and testing.

• Train staff on tools and processes early in the migration to give them the required skillset.

• Build a robust migration factory consisting of people, tools, and processes that help streamline your migration. Operate in an agile fashion increases the velocity of the applications being moved to AWS.

• Assess the application to be migrated and satisfy all prerequisites a few weeks prior to the migration.

• Start small to build experience, find patterns, and create blueprints. Prioritize workloads and run the migration in waves with short migration cycles. Create reusable blueprints for common workload patterns that increase the velocity of the migration. Empower your team to automate the migration steps.

For more detail, see the following:

• Strategy and best practices for AWS large migrations

• A beginners' guide for Finance and Operations teams in their cloud migration journey