Data modeling is a term that has wide ranging applications – it can be used in the context of an exercise as small as displaying two related tables simultaneously as well as a task as gigantic as describing an enterprise model. Hence while buying a data modeling tool,
1. Usage and needs: In projects where a data modeling tool is likely to be needed for a small task, it is better to go in for modeling capabilities available within the database. However, such a tool cannot work in situations where data modeling is a repetitive task and needs to cater to the enterprise level. Hence, going in for evaluation of niche enterprise data modeling tools will help.
2. Reach: In cases where there is a history of projects with data modeling activity, the clientele and size of the data model exercise should be analyzed and then the decision on the reach of the data modeling tool needs to be taken. Hence while going in for data modeling tools, it is important for a company to look at the present needs which are simpler as well as the future reach, which may get complex.
Features: Understanding the business requirements is the first step. The second task is to ferret out the right data modeling tool from myriad offerings by tool vendors making tall claims. It’s key that the company evaluates these niche data modeling tools and then takes a call. Apart from the above mentioned aspects, it’s critical that you check the following aspects that data modeling tools should be able to perform
• UML modeling or use case modeling
• Conceptual, logical and physical data modeling options
• Multi-user and metadata modeling capability
• Ability to capture reporting metadata model and extraction of the physical model
• Reverse engineering from database to PDM (Physical Data Model)
• Modeling lineage
If a company is already using a data modeling tool that has these features, then it’s best to create a forum where the current SMEs can put in their views. The information can be collated and shared, which will help in choosing the right tool.
3. Integration: In case of data modeling tools, it is found that different projects use different tools and hence the data model created will be in a proprietary format as also common format. Hence, data modeling tools should be able to read common format models created by other niche data modeling tools. But this option might not be available in small-scale tools. Backend integration is another key aspect.
4. Cost: The price directly depends on the credibility of the vendor and the tool. Some data modeling tools are either not marketed well due to lack of vision or poor post-sales support.
5. Documentation: When it comes to documenting the modeling activity, it is found that it becomes manual work for the modeler. Hence, data modeling tools should have the capability of producing documentation. In this context, documentation using a data modeling tool requires the following aspects:
• Should be able to provide enterprise information
• Each subject description should be available
When a business subject is exported into the document, all the information about the entities involved, its description, attributes within each entity and how the relationships are set between the entities should be reported both in a business fashion as well as technical. Thus, it needs to be ensured that the data modeling tool is capable of providing documentation to both business-users and technical data modelers. This will cut down a lot of man days needed to create documents.
The user community factor
If there is a high user community for a data modeling tool, resources for the same will be easily available in the market. Hence it is key to survey the percentage of user community available for a given data modeling tool as well as the vendor’s vision and strategies. The best vendors usually come up with an evaluation copy of the respective data modeling tools and allow the user community to download freely.
About the author: Venkat Iyer is the Director and BIM India Lead at Capgemini. He has an experience of more than 16 years in business information management.