Data Migration - Phased Approach
Our organizational approach can be split into well-grained phases:
Data Analysis & Data Quality assessment
Analysis of data fullness and consistency degree. Numeric forecasts on results of data migration. Estimation of systematic data errors, incompleteness. The result of these phase is detailed report listing:- Relevant Data Sources that can be used or should be preferred
- Percentage of err data, suggestions on criticality of errors
- Proposal on Sequence of Data Sources for migration
- Assessment of need and volume of data cleansing
Developing Data Migration Strategy
The strategy includes:- Determining the sequence of actions and iterations
- Planning quality assurance related activities
- Aligning plans with other ongoing projects
- Assigning KPIs to each phase and iteration as data quality and quantity measurements
Designing and Implementing Cleansing procedures
Cleansing procedure is aimed to analyze and remove/mark data that is by given criteria is deemed dirty. This is done by means of automated scripts based on analysis of data model of target system. The result of these phase includes:- Implementation of Cleansing procedure
- Implementation of Reporting on cleansed data
- Specifications and guides of the processes and reports
Designing and Implementing Extract procedures
Extract procedures are intended to extract the data from various source systems with the least transformation into a single format, i.e. Oracle database or MS SQL Server database. The result of these phase includes:- Implementation of Extract procedure
- Specifications and guides of the processes and reports
Designing and Implementing Transform (also called Data Normalization) procedures
All data sources are merged and trasnformed at this step into a single converged intermediate data storage. Maximum data consistency rules are enforced. The result of these phase includes:- Implementation of Data Transform procedure
- Implementation of Reporting on Transformed Data and Errors
- Specifications and guides of the processes and reports
Designing and Implementing Load procedure
The data is finally loaded into the target system. This can be direct load into the target system data storage or loading using API of the target system. The result will be specifications and implementation of the loading process into the target system. Load report, i.e. number of objects loaded, will be provided as well.- Implementation of Data Load procedure
- Implementation of Reporting on Loaded Data and Errors
- Specifications and guides of the processes and reports