This is an obvious and reasonable approach to creating a complex relational data model, and we have found that this technique is well established in its application to life sciences. Experience has shown that in the longer term, however, this approach presents us with scalability and performance issues because it relies on 3 major dependencies:
1.A consistent data set – relational integrity plays an increasingly important role as the rows in the data sets are updated and/or superceded.
2.Speed of loading – how do we take a large, structured data set into another data set with a large number of related tables/indices and, as the dataset grows, maintain an acceptable loading performance?
3.Query performance – to achieve a sufficient level of performance, we must heavily index each table in this structure, simply to reassemble the data into the form that the user will eventually require.
Nested Relational Approach
The Nested Relational approach that GE Healthcare Informatics employs is significantly simpler. Instead of trying to force square pegs into round holes and creating a large conceptual overhead, we will leave some pegs square and allow the round holes to have slightly square-like edges.
In many cases, data must be queried from the private database similarly to how it was found in the original system, but with some additional ”added value” information, such as cross functional references. To do this, we must consider the three issues identified above: we must provide a high performance load and query method that also adds inherent integrity in order to create a highly scalable, added value relational database.
By taking advantage of recent advances in database technology, we can enable a relational database to adopt an