Using the Data Migration Specification, Test Plan, and Test Strategy as your guide:
-
Extract source data
-
Transform source data
-
Migrate transformed data
To successfully perform this step, the implementer needs to work with a business designer possessing an intimate
knowledge of the data that needs to be migrated and a business analyst with knowledge of both the source and target
system data.
Ideally, the completion of data migration should immediately precede the start of production running of the system.
However, when the data is very volatile or very large or when the validation of the migration results is a lengthy
process, amendments to the migrated data might be necessary. This is especially true where manual data has to be
collected, verified, and then entered to the automated system; in these circumstances, data capture might take many
months.
In such cases, you need to plan and define the maintenance procedures to keep the migrated data up-to-date in any
interim period between the completion of the data migration and the initiation of production running. These procedures
need to strike a balance between economy (they will not normally be in use for very long) and accuracy (they must not
introduce errors into the converted data). They should also incorporate controls and audit trails as applicable.
Whenever possible, entry of manually converted data should be performed using the standard data entry processes of the
new system. Occasionally, specially built conversion processes are unavoidable. A typical example is the bulk batch
entry of data which would normally be entered interactively, but where volumes are too high to make interactive entry
practical. The manually migrated data can be identified in the Data Mapping Table contained in the Data Migration
Specification.
|