What is data migration?
Data migration means moving data—often from on-premises locations to a cloud platform.
Defining data migration: What is data migration?
In general, data migration means moving digital information. Transferring that information to a different location, file format, environment, storage system, database, datacenter, or application all fit within the definition of data migration.
To define data migration more specifically:
Data migration is the process of selecting, preparing, extracting, and transforming data and permanently transferring it from one computer storage system to another.
Data migration is a common IT activity. However, data assets may exist in many different states and locations, which makes some migration projects more complex and technically challenging than others. Examples of data assets include:
- Unorganized assortments of files stored across many different devices.
- Applications, operating systems, and environments.
- Relational databases like SQL Server, MySQL, PostgreSQL, and MariaDB.
- Unstructured databases such as MongoDB, Azure Cosmos DB, DocumentDB, Cassandra, Couchbase, HBase, Redis, and Neo4j.
- Data lakes, data blobs, and entire datacenters.
As a result, data migration projects require planning, implementation, and validation to ensure their success. Learn more about cloud migration and other types of migration, here.
Planning a data migration
Before even beginning to gather requirements for and scope a cloud data migration, organizations need to start by discovering and assessing what data they actually have. They must map the data—find out how much of it there is, how diverse it is, and what quality or condition it is in.
They'll likewise assess the impact of the migration on the organization, establish who the stakeholders are and who has relevant expertise, assign responsibilities, set budget and timelines, and agree on how everyone will communicate about the data migration project.
After scoping the project, teams then design the migration, which includes selecting data migration software and hardware they'll use when they move the data, creating specifications for the data migration, and determining the rate at which they will migrate the data: all at once, just a little bit at a time, or anywhere in between. Many organizations seek help and guidance right-sizing their migration—especially when moving to the cloud.
Implementing a data migration
When planning is complete and the migration is designed, teams begin implementation. They build the data migration solution according to the requirements and step-by-step migration guidance set forth in the planning phase and begin transferring the data.
As the data migrates, teams monitor and test it to ensure the data is transferring properly and free of conflicts, data quality problems, duplicates, and anomalies. This monitoring and testing take place in an environment that mirrors the production environment and enables teams to quickly identify and remediate any issues with the data migration.
Validating a data migration
After all the data has been migrated and implementation is complete, teams will audit the data in its new configuration and validate that the data has been transferred accurately. Teams take the old data configuration out of service only after the data migration is validated by technical and business stakeholders as well as anyone else—including customers—who might use the data.
An organization may need to or choose to migrate data for many different reasons. At a high level, these reasons can include reducing costs, enabling innovation, increasing performance, creating higher availability, and strengthening security. As organizations make the decision to migrate data, they'll need to consider the integrity of the data, the cost of migrating, and the impact to the business and its customers.
Some specific scenarios and business cases that may require data migration include:
-
Upgrading or replacing legacy hardware or software so the organization can meet its performance requirements or be more competitive.
-
Lessening environmental impact—and decreasing operational costs—by moving to a system that has a smaller footprint and uses less energy.
-
Reducing or eliminating the expense of hosting the data in on-premises datacenters by migrating to the cloud.
-
Centralizing data to enable and facilitate interoperability or relocating to a more secure datacenter.
-
Backing up data to allow the organization to better prepare for and execute disaster recovery.
-
Organizations that need to migrate data as part of a modernization effort often seek expert advice and help setting up their cloud environment and guiding their cloud data migration end-to-end. Learn more about the benefits of cloud migration.
Data migration vs. data conversion: What's the difference?
To have a clearer understanding what data migration means, it's important to know what data conversion is and how it relates to data migration. Often, there is confusion around whether an activity or project is data conversion vs. data migration because by definition, data migration includes data conversion. However, data conversion is just one aspect of data migration, so the two terms cannot be used synonyms for each other.
Data migration means moving data from one place to another, whereas data conversion means transforming data from one format to another. The following comparison highlights more of the differences and similarities between data migration and data conversion.
| Data migration | Data conversion |
|---|---|
| The data is moved to a new datacenter, location, system, or environment. | The data is moved to a new application. The datacenter, system, or environment may remain the same. |
| The format of the data may remain the same. | The format of the data is transformed. |
| The process consists of planning, implementation, and validation. | The process consists of extraction, transformation, and loading. |
| Data migration often includes data conversion, but data conversion is not always required. | Data conversion is often one of the first steps in data migration, but data migration can happen without data conversion. |
Data migration can introduce new applications that read information in a way that is different from how the legacy applications read information. In order for the data that worked with the legacy application to work with the new application, teams must transform the data into a format that the new system can understand and use. That transformation process is data conversion. Through data conversion, teams can move data from a legacy application to either an entirely different application or a different version of the same application. The data is extracted from the source, transformed into a new format, and loaded into the new application.
Often, carrying out a successful data migration means that teams need to transform the data, which happens early in the data migration process, before the data is moved to a new place. Data conversion does not include profiling, cleaning, validating, or—after the data has been moved—performing quality assurance tests on data. So it becomes less a question of data migration vs. data conversion and more a question of how a project includes data migration and data conversion.
Types of data migration
While every data migration project is different according to the systems and data involved as well as the organization’s objectives, data migration can be classified into these five broad categories:
These are not the only types of data migration, and a data migration project can include several types of data migration. For example, if an organization decides to move data from an on-premises server to a server operated by a cloud provider, that project might constitute a cloud migration and a database migration. The five categories are helpful because they provide a general outline of data migration scenarios and the reasons an organization may undertake that particular type of data migration.
Storage migration
Storage migrations are the most basic types of data migration, fitting the literal definition of data migration. These migrations consist of moving data from one storage device to a new or different storage device. That device can be in the same building or in a different datacenter that's far away. The device may also be of a different kind, such as moving from a hard disk drive to a solid-state drive. Migrating data to the cloud or from one cloud provider to another is also a kind of storage migration, though the specifics of those types of data migration are better understood as cloud migrations.
Organizations may choose to do a storage migration when they find the need to upgrade their equipment or infrastructure to achieve faster performance or save money on scaling. The new technology may also enable the organization to manage, secure, back up, or recover data more effectively. During a storage migration, organizations also have the opportunity to clean and validate the data, though it is less often that organizations opt to change the format of the data during this type of data migration.
Database migration
This type of data migration often requires data conversion because database migrations typically involve moving large amounts of data to an updated or different database engine or database management system. Database migrations are more complex than storage migrations because not only is more data being transferred, but that data is likely changing in format too.
Database migrations may become necessary for organizations when they need to upgrade their database software, migrate a database to the cloud, or change database vendors. Before migration begins, teams must ensure there's proper capacity for the database and test to make sure that there will be no impact to the applications that use the database.
Application migration
An application migration involves moving data to a new computing environment. This type of data migration is an example of a data migration that combines several others. Migrating an application may require both database migrations and storage migrations. The database that the application uses will need to be relocated—sometimes even modified in format to fit a new data model via data conversion—along with the files and directory structure the application requires to install and run.
Organizations may carry out an application migration when there is a change in the software the organization uses to perform a business function, the vendor that provides the software, or the platform where the software resides.
Cloud migration
Much like two other types of data migration—storage migration and application migration—this type of data migration involves moving data or applications. The key aspect is that cloud data migration refers specifically to transferring data or applications from a private, on-premises datacenter to the cloud or from one cloud environment to another. The extent of the migration will vary. A cloud migration may involve moving all data, applications, and services to the cloud, or it may entail moving just a select few to meet a strategic purpose or business need.
Migrating to the cloud enables organizations to scale with fewer limitations, provision resources more readily, upgrade with less fiction, spend more effectively, and innovate more rapidly. With their data and applications residing in the cloud, those organizations are no longer required to maintain the machines and infrastructure that were storing those assets on-premises.
Business process migration
This data migration type refers to moving data and applications in order to better manage or operate the business itself. In a business process migration, the organization may transfer any kind of data—including databases and applications—that serves products, customer experiences, operations, practices.
Organizations may undertake this type of data migration to optimize or reorganize how the business is run, to better compete in the market, to offer a new product or service, or to complete a merger or acquisition.
Data migration tools
In order to carry out their migration, teams will use various data migration tools to move the data and modify it as needed. Some teams will choose to build their own data migration tools from the ground up. The advantage of building data migration tools is that teams can tailor the tools to their specific systems and uses. However, coding data migration software can take a lot of time, require a lot of manual integration and re-implementation work, and incur costs that may be better spent on other parts of the data migration process. Self-scripted data migration tools may also run into challenges scaling or handling many sources of input.
Instead, teams may opt to use existing data migration software to make the act of moving data simpler, faster, and more efficient. Often, software specializes in helping with a particular kind of migration—like moving a SQL Server database to the cloud. But even with the software, the team still needs to know all about the data they're moving, how much they'll migrate and when, what changes they'll need to make to it, and if there are any issues to resolve once the transfer is complete. And those teams will also need to choose between on-premises data migration tools and cloud data migration tools.
Which type of data migration software to use
Teams can choose from on-premises, cloud-based, or self-scripted data migration software. Generally, on-premises tools work well when the data and target systems are all on-site and within the same organization, cloud-based tools are best when moving different data systems or replatforming to the cloud, and self-scripted tools can be good for small and highly specific projects. However, because data migration projects are complex, there are many more factors to consider in choosing from the different types of data migration software available. This chart suggests which tools excel depending on the features of a given migration scenario.
| Self-scripted tools | On-premises tools | Cloud-based tools | |
|---|---|---|---|
| Data volume and type | | | |
| Small amount |  | | |
| Large amount | | | |
| Supported format | | | |
| Unsupported format | | | |
| Source and destination | | | |
| Single-site sources | | | |
| Multi-site sources | | | |
| Cloud destination | | | |
| On-premises destination | | | |
| Common source and destination | | | |
| Uncommon source and destination | | | |
| Project needs | | | |
| Scaling required | | | |
| Scaling not required | | | |
| Control of storage devices | | | |
| Local access | | | |
| Global access | | | |
| Compute and storage on demand | | | |
| High uptime and reliability | | | |
How to choose a data migration tool
- In addition to the criteria outlined above, teams and organizations will consider other factors in selecting their data migration solution. Those factors include:
- Budget and timeline
- Expertise and experience of the team.
- How much scale and flexibility the organization needs
- Relationship with the provider of the data migration tool
- Security and regulatory compliance
- Uptime or other SLAs
- Potential impact
- The users of the data
- Operating systems
How to start cloud data migration
Once an organization is ready to consider data migration, they might begin exploring their options for data migration tools or a data migration partner. To learn about the advantages and process behind migrating to Azure, explore these resources:
Frequently asked questions about data migration
-
Data migration is the moving of digital information. Transferring that information to a different location, file format, environment, storage system, database, datacenter, or application all fit within the definition of data migration.
-
Data migration means moving data from one place to another, whereas data conversion means transforming data from one format to another. Data conversion sometimes takes place during data migration.
-
The types of data migration can be classified into five broad categories: storage migration, database migration, application migration, cloud migration, and business process migration.