DATA MIGRATION
The process known as data migration is, in its most general meaning, the transfer of data between formats, types of storage and computer systems. Usually it is performed when a new system is adopted (through change, upgrade or purchase) and the data in both systems need to map to each other without difficulties. As the very process of data migration requires a considerable amount of time and labor, in most cases it is done automatically, although manual methods are also frequently used. These are: manual data entering, moving files between folders and disks, developing custom software and more.
The method chosen for a given system is totally dependent on both the system involved and the state of the data supposed to be migrated.
It is common that data cleansing is performed during the procedure of data migration for the purpose of data quality improvement and matching the requirements of the new system.
There are two variables to be taken into consideration when deciding on data migration - data volume and data value. Although they are both important for the process, it is the data volume that is the most crucial criterion once the decision of data migration has been made.
- Before the proper migration begins, there is a set of analyses to be performed:
- the analysis and definition of source structure
- the analysis and definition of target structure
- mapping between source and target structure
- definition of the migration method.
Following all these steps every time the data migration is required, may help to save actual costs of this not simplistic and rather expensive task.
For business purposes, data may be migrated to meet and to depend on various requirements and on this basis four main migration categories have been proposed:
- storage migration (virtual migration - minimal impact on data format and content)
- database migration (physical migration to prevent significant alterations in the database)
- application migration (data migration handled by application programming interfaces - integrity protection)
- business process migration (movement of data between two databases to ilustrate specific changes).
The first two categories are routine and significantly less complex than the last two, which cannot be taken care of with involvement of only the IT department, since they affect the users of applications and processes directly.
It is important to make a distinction between data migration and data integration. Data migration is rather a project where data is transferred from one set to another and sometimes eliminated from the source, whereas data integration is a permanent process, the purpose of which is the responsibility for the data flow between numerous data stores and applications. Also, it is worth mentioning that although standard ETL technologies are associated with both actions, they are more likely to fit in the category of data integration.
Apart from its unquestionable usefulness in business environment, data migration can be an excellent approach to digital preservation of objects (eg. reproduction of paper documents onto microfilms).
Although data migration is neither easy nor a cheap process, the benefits of it are worth all the costs and this fact is understood by all those who work with data on a daily basis.
Learn more