Data migration for the version change of the central CUSA Prevention in C5 technology
In its capacity as the statutory accident insurer, the Bayerischer Gemeindeunfallversicherungsverband (Bayer. GUVV, LUK) is responsible for accident prevention, as well as rehabilitation and compensation for employees in claim cases in Bavaria. More than 75,000 companies and institutions came under its aegis in 2009. CUSA, the business solution of HDP GmbH for German statuary accident insurance, serves as the core system for automated processing of the relevant procedures.
Extensive data migration was required in the course of the transformation of the software forming the foundation of CUSA Prevention, the procedure for technological field service in 2011. This coincided with a change in database from Oracle to db2 i5. Drawing on its in-depth data migration competence and the use of suitable tools, an expert team from mgm technology partners worked with the client team to successfully provide the new system with the data – on time and on budget.
The task: Planning the transformation procedures and data loads based on the existing migration concept
The new system had to be loaded from two data sources in order to put the current version C5 of CUSA Prevention into production. The predecessor version C4 delivered the ongoing data from the day-to-day business (such as documentation of the customer support reports). In addition to that it was necessary to fill the new CUSA installation with stock data (such as members, performance areas), which was on hand in the data management form customary for the host. Both data fillings required a technical transformation (ETL) as well as coordination of the data contents with the Bayerischer GUVV organisation.
Flexible integration of project members in a dynamic transformation process:
- Coordination of the data preparation with the responsible departments from the source systems
- Ensuring the correctness of the data use in CUSA Prevention with respect to business
- Tool selection (Kettle ETL tool from the Penthaho Suite (OSS)) for technical depiction
of the transformation processes
- Iterative performance of the data loads and optimisation of the transformation rules
- Introduction of an automated data reloading procedure for changes in the stock data
- Quality assurance of the provided data with the departments
“Thanks to its expert background, the mgm team was right at home in our world. They were sound expert partners in the coordination work necessary for the data migration and succeeded in giving us an eye-to-eye understanding of the technology employed,” explains Jürgen Renfer, Department Head at Bayerischen GUVV.
Skillful data transformation and the use of Kettle made the technical migration efficient
The team chose Kettle, the ETL tool from the Pentaho product family, as the technical tool for the data transformation.”Kettle allowed us to address the existing Oracle and db2 i5 DB systems directly via existing connectors,” explains Robert Popp, migration expert at mgm. He designed a one-step transformation and loading process based on Kettle that used XML for the technical preparation of some data. This was necessary because the business data was stored in Blobs. Only a conversion of the Blob data to Clobs allowed its usage for further processing.
In the end, it was possible to migrate to the target system in one step without the customary staging area for precautionary consolidation and data appraisal. “Using our procedure and Kettle, the project team was able to show the experts at Bayer. GUVV their data and existing inconsistencies directly and to make changes immediately and traceably without programming,” Popp adds as the advantage of his procedure.
On time and on budget project completion and cooperation at eye level ensure customer satisfaction
In order to comply with the time frame for project completion, it was critical to the mission that the users could work on removing data inconsistencies that typically occur during the initial data loading as soon as possible.
Thanks to the selected migration approach, only four (loading) iterations, including the accompanying coordination with business experts, were required until execution could be confirmed: Error-free and complete (100%) data takeover into the production system. That the, to some extent, very detailed coordination concerning data corrections could also be completed so swiftly speaks for the project management’s activity tracking, but shows above all the great spirit within the whole team.
*This reference has been officially confirmed by the Bayerischer GUVV.