Diverse data structures and formats, representing the same underlying information in different ways, are a fact of life. Integrating these applications and data structures is a huge IT cost driver.
To analyse how different data sources relate to each other, and to integrate them, people commonly develop mappings between them. But the common approaches to mapping, as supported by many commercial integration tools, are flawed and weak. Why? Because they use direct structure-to-structure mappings – so that each mapping inherits all the physical complexities of both mapped data structures. This makes them complex, fragile, and hard to create and maintain.
Rather than map data structures direct onto each other, we map all data structures onto a common logical (UML) model of the information. The UML model, unlike the data structures, has no extra physical data complexities. It tells you what the data mean, with no complications. This gives mappings which are simpler, more meaningful, easier to create and maintain. Model-based mappings are the foundation for robust and cost-effective integration of diverse data sources.
For further Information Contact us