Creating Single or Multiple Collections
One of the key capabilities of Watson™ Explorer Engine platform applications is their ability to search data that is stored in many different applications and locations. Because the amount of data stored in enterprise applications such as email servers and archiving systems, content management systems, databases, and so on, can be quite large, the corresponding search collection for that data may also be quite large.
When planning a Watson Explorer Engine platform application, a common question is whether you should create a single search collection that indexes all of the data from a single data repository, or whether you should create multiple search collections for subsets of that data. Unfortunately, there is no single, generic answer to this question because the data on which you must base this decision is too site-dependent.
To help you determine how to organize the collection(s) that will be used by your Watson Explorer Engine platform application, the following sections discuss the basic ways of structuring search collections, and then highlight the positive and negative aspects of each.