Solution

For a company suffering from poor performance, including lack of flexibility and scalability, the path to the cloud is often the preferred way out. In b.telligent's experience, pure migration often fails to achieve the desired result in terms of operation and use of scalable services, for example. To be able to take all of medi's requirements into account and implement a long-term solution for the client, b.telligent developed the appropriate architecture jointly with the client on the basis of guiding principles, relying on a cloud-first approach in this process. Certain aspects of the platform which would particularly benefit from the cloud were redesigned. On the other hand, developments which had already required a great deal of work were to be migrated.

Specifically, all 17 source systems at that time were connected via the Azure data factory. After development of a zone concept for the data lake, the data were processed in this lake and thenceforth represented the linchpin of the new, modern data platform. In contrast, the actual data warehouse with its data vault modelling was migrated by means of lift-and-shift to the PaaS service known as SQL managed instance. This approach proved to be absolutely correct. The business logic's complexity became clear in the course of the project. New development from scratch would certainly have taken years. However, migration of the tabular cubes to the cloud was completed within a few days, thanks to 100% compatibility. The platform's finishing touch was a complete transfer of existent SSRS reports into the Power BI world in order to combine standard and ad-hoc reports under a single interface in future. Introduction of a standardized deployment process now makes it possible to release new features individually for the platform.

Last but not least, b.telligent encourages sustainable use by clients themselves in all projects. That is why an important component of this project was to successfully impart the necessary skills to the colleagues at medi using the principle of training on the job. "Close cooperation between colleagues from medi and b.telligent was particularly satisfying for us. Due to the pandemic, much of the project took place remotely, this certainly posing a number of hurdles. Thanks to excellent cooperation by the mixed team and good project management though, these hurdles could be overcome and the project completed in time, within budget and at the required quality," says Bastian Dickehage, head of analytics at medi.