Architecture

Our answer to DWH reference architectures

Having a datawarehouse architecture is important because you need a sound strategy on what to do with the data. You maybe also need help picking platforms and architectures for managing data. Therefore, we are going to promote the most used architectures as an out of the box solution and providing guidance when a certain reference architecture should be used and how it can be built within our developer studio

Kimball - dimensional modelling

Create an organization wide or department specific datamart based on the dimensional modelling approach. It provides the capability for reporting & data analytics. The datamart contains a simplified structure that is more intuitive for data analysts to discover and query. It stores data in a way that maximizes the performance of BI tooling. Creating a simplified model does not mean designing and building a datamart is simple this because first a lot of backroom data logistics needs to be designed and implemented.

Hub and spoke

Create a “go-to” place for data within your organization. With a data hub point-to-point connections between data suppliers and data consumers are replaced by a centralized data store. The Data Hub can negotiate deliverables and schedules with various data consuming parties. Data can be served in multiple formats that do not have to be integrated. It can serve both operational data (ODS) and historical data.

Inmon - corporate information factory

If you want to go a step further than a Data Hub and also need to harmonize and historize data in your organization: the Enterprise Data Warehouse (EDW) can solve naming, structural and semantical differences between data sources and data consumers. Create an integrated view on data by combining data from multiple sources. Add value to data such as de-duplication, quality and reconciliation checks.