The decisive factors for data modeling solutions depend upon complexity of data, frequency of execution and extent of usage that the tool will deal with. In this context, it goes without saying that the information provided by vendors is flavored to portray their solutions in better light. Keeping these in mind, other factors for evaluation of data modeling solutions include the following:
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
1) Size of the company and complexity of data: Your organization’s size is a key factor to be kept in mind while selecting a data modeling solution. In an architectural environment, the dependence on the data modeling tool is less, due to a higher reliability on the programmer’s intelligence.
Large databases will require an advanced tool. A new data modeling solution should be initially set using a simple tool and then upgrades. This is a precaution since a company might want to migrate to a larger database in the future.
Besides the well-known names, the smaller, unknown tools should be looked at to fit their features to the particular requirement of the company.
2) Organizational factors: These include the existing software, operating system, migration effort and timeframe. You should also utilize methods like scoring and questionnaires (with operators who produce high-levels of data) for proper evaluation of the data modeling solution.
3) Compatibility of the tool and database: It helps if the data modeling software is compatible with database. It should be able to identify whether it is an attribute or an entity, thus identifying the entities and the relation between the tables. Also identify mismatched data and offer an advanced warning. Data merging is not a widely available feature, but must be kept in mind if several formats are going to be dealt with in the data model.
4) Team based modeling capabilities: Leading standards have credibility from large clients that use their data modeling solution. The benefit is that a modeler can work on a parallel model and also offer user sub-models. This makes integration of data simpler.
5) Historical models: This feature offers the benefit of tracking changes in the software development. The data modeling solution can thus be modified and amended to the software being developed.
6) Advertising of tools: Besides the well-known names, the smaller, unknown tools should be looked at to fit their features to the particular requirement of the company. Reference and cross checking with organizations that have such data modeling solutions in place is advisable.
7) Version capabilities: The data modeling solution should have version capability to identify changes that happen at various stages of development and reviewing.
8) Virtual modeling: Repositories in BI tools use virtual modeling. This is also useful in storing metadata. The old model is used to develop programs. The repository will enable to convert this to a new model. Large databases could be converted to virtual models and then used. The data modeling solution should be capable of integrating this data.
About the author: Nitin Gadhe is a COGNOS certified, data warehousing and business intelligence consultant at Cybage, with experience in design, implementation and support of Global Cognos Business Intelligence Solutions.
(As told to Sharon D’souza.)