Home > Documentation > Miscellaneous

How to Migrate Database into Cloud

Migrating to cloud services has become a long-term trend in the world of databases. The main three reason of growing popularity of "Database as a Service" model are: reduced TCO (total cost of ownership), easy scalable hardware and software architectures, virtualization of data.

Most vendors of popular databases systems offer cloud-based solution as well as large IT corporations:
  • Azure SQL - cloud service for Microsoft SQL databases
  • Oracle Cloud - "Database as a Service" solutions for Oracle databases
  • Google Cloud SQL - web service to create, configure and use relational databases based on MySQL engine
  • IBM SmartCloud - IBM cloud service for DB2 databases
  • Amazon Relational Database Service (RDS) - cloud services for IBM DB, MySQL, Microsoft SQL and Oracle databases by Amazon

Basically, the process of complete migration from database server to the cloud can be represented as sequence of following steps.

  

1. Plan and setup. A key part of this phase is doing detailed research of features that are not supported in the destination cloud and the necessary application changes caused by migration.

This step requires most part of human resources because it is connected with understanding of database applications semantic. Even though there are some tools that can help identify missing features in the destination database cloud, the project manager must understand the role of that functionality in the particular database system and find design alternatives. The decision to proceed with migration must be based on estimation of development efforts and cost of deploying the alternatives, rather than simply using missing features as reason to not migrate to the cloud at all.

2. Development. On this phase people assigned to development should perform tasks generated from the planning and design. Most databases migrated from another systems require some application changes. But even inside one database system, cloud solution can have minor differences compared to database server. For example, when migrating data from Microsoft SQL to Azure SQL it is necessary to create clustered indexes for each table that does not have it. Also, database developer must remember that Azure SQL does not support RESTORE statement and cannot attach databases to the SQL server.

Even if no database objects are changed during the migration, the application architecture will probably require modifications due to different security, logging, error handling and robust retry logic.

3. Test. This is probably the most important phase of the entire migration process. People responsible for testing should recognize that it is far less risky to validate regularly and fix errors being found fast than having code without validating for long periods.

Below is the list of testing areas deserved most attention of both developers and planers as basis for performance and functionality tests:

During the test phase database activity should be isolated, otherwise it may distort results of benchmark.

Intelligent Converters Solutions

Intelligent Converters provides a range of software and services to migrate data from any sources (MySQL, MS Access, Oracle, Postgres, DBase/FoxPro, MS Excel, Microsoft SQL, IBM DB2) into private or public cloud. Our products work with all popular cloud-based services: Azure SQL, Oracle Cloud, Google Cloud SQL and Amazon Relational Database Service. The appropriate converter will migrate the data, constraints, indexes and queries (views) into the cloud storage with required transformations. And our experts will carefully guide through all steps of migration process listed above providing necessary information and advises.

This combination of high performance migration tools and cutting-edge skills allows Intelligent Converters to provide unique solutions that have helped hundreds of customers to migrate databases to both private and public clouds.

See also

How to connect to Azure SQL
How to connect to cleardb MySQL