The easier part is certainly whether the DWA tool implements Data Vault correctly. Here there are criteria to be met. On this website you will find 13 of these criteria and requirements for Business Vault and Data Mart. If there are your own additional cherished patterns or special requirements such as bitemporal histories, these still need to be added by you.
Then there are the expectations for functions that are not directly involved in the construction of the data warehouse. The following overarching functionalities are needed:
- Data Lineage
A representation of the dependencies in data preparation. Either at attribute level or at entity level. This makes it possible to understand how changes to the interface affect the key figures or from which data sources a key figure is composed.
The data for the lineage should not have to be documented separately but should essentially result from the implementation and be maintainable with little effort in the course of further development.
- Error Handling
Handling of errors and notification that an error has occurred. Data Vault also loads erroneous data, so errors should be reported in an ‘error mart’ or similar.
Determine the sequence of the individual processing steps, taking into account the interdependencies between them.
The control of the preparation processes with the possibility of parallelising the loading and executing it time-controlled. Ideally, load balancing
can be carried out.
The distribution of the generated code to development, integration and production. The support of tests.
- Supported databases
- Prerequisites for the installation
Some of these functionalities can or must be implemented with additional tools. Here it is important to see which tools are needed and how they fit into the existing organisation.
The considerations for fitting the tool into one’s own organisation and development environment are highly individual. Here is a list of possible candidates:
- Coding skills of the BI team
- Requirements for individual customisations
g. company-specific solutions that can be automated
in the Business Vault or Data Mart
- Supported databases
depending on existing DB know-how
in the company as well as the possibility to break new ground
- Cloud strategy of the company
including the question on-premises or SaaS)
- Integration with existing scheduler
- Embedding in existing CI/CD solutions
The criteria for Data Vault and the overarching functions are recorded and explained within the framework of the DDVUG DWA Challenge. The considerations for embedding in one’s own organisation flow into an evaluation system:
- Are there criteria that simply have to be met?
For this, give
a clear yes/no and if no, exclude the tool further on.
- Are there criteria that are more important than others?
If so, how many classes of them?
For example, if there are three different classes: important, less important and not important.
Then the fulfilment of the not important criteria gets 5 points each, the less important 10 and the important 15 points.
Now give points per tool and criterion for the degree of fulfilment and add them up. Who has the most points? And is the price right in relation to the second place winner?
The procedure described here is a very simple one. However, there may also be requirements from your own purchasing department that have to be followed.
If in doubt, don’t think too long and just do it. You can always adjust it afterwards…