Data Migration

DEFINITION of 'Data Migration'

Data migration is the process of moving stored digital information between computers, systems, or formats. Data migration occurs for a number of reasons, including server replacement or maintenance, a change of data centers, data consolidation projects and system upgrades. As much of a company’s corporate knowledge and business intelligence is contained in its data, any data migration project must be done carefully to minimize risks.

BREAKING DOWN 'Data Migration'

Data migration represents a significant risk to business continuity if it is not done properly. Loss of data is, of course, a worse case scenario, but companies must also deal with downtime, compatibility issues, and overall system performance issues. Data migration is made more difficult by large amounts of data, diverse formats of data, and divergent data habits within a corporation.

To minimize the risks of data migration, companies create detailed data migration policies that prioritize back ups, moving order and concurrent data environments when possible. If it is not possible for a company to run a pre-migration environment while the new environment is prepped, then there will be significant downtime as business operations on the current applications are suspended to allow for the data migration. This type of stop, transfer, and start data migration may be required when moving to new platforms or when there are hard limits on physical storage and swaps or fixes are needed on the existing storage technology.

Zero Downtime Data Migration

The zero downtime migration model depends on having enough storage to create and run two complete environments. A full copy of a company’s data is taken into the new environment and tested while employees stay in the old environment. The bugs are worked out of the new system, ensuring that all the applications still work and everything is where it should be. After testing is complete, a fresh copy is brought in and all the employees are switched over to the new environment. The old data environment is sometimes left open for a period of months so that employees can get files from the old data system but not write new data to those servers. In all data migrations, a post migration data audit is done to check for data loss.    

One thing that can improve a data migration is to clean out and standardize data practices prior to migration. The organization of a company’s data is often a reflection of the different filing habits of its people. Two people with the same role may well use completely different practices. For example, saving contracts by vendor in one case and by fiscal year and month in another. Unifying data practices can be a much bigger job than actually migrating the data, but clean, coherently organized data backed up by clear policies helps to future proof a company’s data for many migrations yet to come.