You are on page 1of 12

Data Migration

Method outline
A successful Data Migration is defined by a set of quality
criteria
• Agreed stakeholder impact – customer, organization & third parties

• Reliable and correct execution – supported by a transparent validation process

• Controlled migration process – with well defined decision points

• Auditable – process is fully reconstructable

• Right cost and duration – adjusted to risk appetite

A seasoned method balances


attention to the criteria

2
2
Data Migration is represented as a Workstream in a
transition program
1 2 3 4
Analysis Design, Build & Component Test Integration Test Execute

Commercial Customer Commercial Migration Process Customer


Impact analysis Client Communication (i.e. letters) panel
Preparation

Data Data Data Quality Data Quality


Quality Analysis Improvement Improvement

Realization of target application New objects in


Target Application Ist/Soll landscape Test of target landscape for Target Applications
Gap Configuration of Target Applications new Objects (Commercial Go
Landscape
Analysis (i.e. products, …) Live) Pilot
Product/Chann Work Process Definition and Dry-Run of
Work Processes and el matrix Description Work
Instructions Work Instructions (i.e. Aris) Processes

Data Migration Functional Migration Process Migration Dress Pilot


Data Strategy (incl. Data Mapping Tests Rehearsal
Migration Acceptance Design of Migration Components Migration
Criteria) Migration System (Components) Business Process Tests Iterations

3
3
Data Migration done Agile: series of incremental
migration iterations
Gap analysis
Number and duration of iterations depends on:
• Desired level of automation
Data analysis • Available IT capacity
• Available capacity for manual migration processes
• acceptabele customer timpact & impact on own organization
Migratie
strategy

Migration Tranche 1

Migration Tranche 2

Migration Tranche N

Execution
(optional)

Close out

4
4
Migration Iterations consist of a set of standard activities
Objectives, scope &
acceptance criteria

Data Analysis

• Data Ware House Design, Build, Test


Configuration and adaptations of tagret applications

Data Cleansing

Data Mapping

Design, Build, Test


Extract, Transform, Load, Validate

Migration test
Business process test
• test cycles (2) Execution
• Dress Rehearsal
(0, 1 or n tranches)
• Internal Tranche
• Live Migration

5
5
Migration activities result in standard migration
deliverables (with dependencies)
Program Production Commercial Clean Readiness
Objectives Data Migration Production Assessment
& Scope Approach Data
Ist/Soll Target Work
Gap Application Instructions
Analysis Landscape

Data
Mapping Migration Business
Data Test Plan Process
Analysis Test
(source) Dress
Component Migration
Design System Reheasal Execution
Component
Migration
Design
System/Int.
Migration Functional Technical Test
Strategy Migration Migration Design Migration Status/
Process Process Test Report Evaluation
Run Report
Book

Execution
Governance

6
6
Data Migration Strategy defines the outlines of the migration
Subjects to be addressed

Migrating Objects (incl. volumes)

Source & Target Systems

Scope (i.e. active objects vs historic data)

Stake Holders

Acceptance Criteria

Migration Approach
• big bang vs iterations, slicing criteria,
• On-/Off-Line/Real time/Batch
• Timing – critical migration windows vs. synchronization source/target
• Load approach (database, application, service)
• Test approach (Business Process Tests)
• Manual/Automated
• Pipeline issues
• Rollback / fallback
Validation Approach (i.e. confidence checks)

Customer Impact, systems availability, communication

Data Quality (required vs desirable)

7
7
Data Analysis provide insight in production data volumes,
relations and frequency distribution
Excel Pivot Table SQL Queries Data Analysis Report

The truth is in the data

8
8
Data Migration designed according to Extract Transform Load
Validate pattern
Each table contains source, timestamp and where Counts and Sums are calculated at each point where
applicable reference to transformation rules data is handed over to next stage

Conversion Engine
Source Target
Systems Systems

Landing Staging Transform Load

Validation Tool

Extracted Validation Extracted


Files Database Files

Validation Tool Control Model

 Validation tool shares platform with Conversion Engine  The control model defines where the checkpoints in the migration process
 Both the engine and the validation tool are based on extraction files from are created and the metrics
Source and Target  Every time data flows from one stage to the other, a report is generated to
 Mapping is conducted independently from conversion engine to ensure show counts, sums and cross connections
proper validation

9
9
First Test Cycle serves as proof of concept and learning
curve, subsequent Test Cycles iron out defects
Cycle #1 Cycle #2 Cycle #3

Set-up preconditions
Execute migration
Conduct BP Testing
Wrap-up results

Fix defects

Proof of concept Test completion


• Goal 1st test cycle is a complete run to proof the concept • Goal of subsequent cycles is to complete test execution until defects are
• Duration needed for sub-sequent cycles can be derived removed
• Both full migration execution as well as business processes in • Every cycle will be based on a new release of migration engine (except the last
target in scope one which must be based on a stable version)
• Duration of a cycle is not expected to decrease significant

10
10
Successful Data Migration execution relies on five
deliverables

Five Deliverables for Data Migration Execution Stage:


1. Governance process and structure with representation of key stake holders
2. Central control room capable of directing the migration process
3. Run book with full details of data migration activities (who, when, where and dependencies)
4. Proven migration system and process – passed successful Dress Rehearsal
5. Readiness assessment artifacts

11
11
Successful Data Migration requires skills and experience

DataDogs have Data in their DNA


20 Highly trained, skilled and experienced professionals
More than 15 successful migration projects

Vincent Wormer Jacco de Gooijer

Datadogs is the right partner for a


successful Data Migration
Ruud Kapteijn Manuel Noechel

12
12

You might also like