Artifact: Data Migration Specification
This artifact describes data sources to be migrated, and specifies the mapping as well as any required data transformation processes between the source data and the target data.
Domains: Development
Purpose
This artifact is used to document how source data is transformed into target data, either as one-time data conversions or via ongoing data interfaces. It is also used as a basis for validating the data mappings between systems.
Relationships
Description
Main Description

This artifact is a detailed description of the requirements for accomplishing data transfers. A data reference can be either a source (sending data into the system -inbound) or a target (receiving data from the system -outbound).

Appropriate data sources are found for each of the target data elements, and then rules are drafted to extract, transform, and move the source to the target data element. Rules are developed for the following:

  • Data selection - rules describing the criteria by which data is to be selected
  • Data cleansing - rules describing how the source data will be cleansed as it is extracted
  • Error detection and correction - rules defining how missing or invalid data should be treated
  • Data mapping - rules describing what must be done to the source data (transformation) to satisfy target data requirements
Brief Outline

This artifact contains the following information:

  • Scope of the Migration
  • Data Profile
    • Data references (sources and targets)
    • Data attributes
    • Data attribute dependencies
  • Data Mapping Table
  • Data Transformation Processes
Illustrations
Key Considerations
  • It is preferable to have legacy developers from the source data reference as participants in the data mapping process.
  • Data mapping rules may be stored in an automated tool, preferably one with a repository.
  • The logical level of the data model should be completed, and the external data references identified before the data mapping is begun.
Tailoring
Impact of not having

If this artifact is not created, the formal rules and procedures for data extraction, cleansing, and subsequent transformation are not documented for review.

Data is often a critical component of any system; therefore failure to analyze and correctly map data between different data structures or systems may lead to unexpected results.

Without this artifact, developers are required to have an understanding of the underlying data and perform the data mappings directly in the code.



Reasons for not needing

Some reasons for not needing this artifact are:

  • There is no data to be migrated.
  • The data migration is straightforward enough that it can be directly implemented and verified without formal up-front analysis.