IBM E02HMLL-I Implementation Guide - Page 157

Chapter 8. Configuring database connection pools At run-time, it takes time for processes to establish new connections to the database. You can minimize that time by establishing database connection pools in advance for use by your collaboration and map processes. Database connection pools can improve performance and enable direct database access from within a collaboration. A database connection pool can be used by multiple collaborations and maps, and each collaboration or map can use multiple database connection pools. This chapter contains the following sections: v "When to use database connection pools" v "When not to use database connection pools" on page 147 v "Creating database connection pools and database connections" on page 148 v "Validating database connection pools" on page 151 v "Modifying database connection pools" on page 151 v "Using database connection pools in collaborations and maps" on page 153 v "Configuring transaction bracketing" on page 153 When to use database connection pools This section describes some of the situations in which you might want to use database connection pools. Performing routing The logic of a business process may dictate that a business object must be routed to different destination applications depending on the value in one or more fields in the business object. For instance, a site might store and process customer entities in different applications depending on the value in an attribute such as CustomerType. The collaboration template would need to retrieve and evaluate the value in that attribute and make a decision as to which destination application to send the business object to based on the value. Although this can be accomplished with control flow structures in Java, the pairings between the values and destination applications are then hard-coded in the collaboration template. If they need to be changed because of a change in procedure, or added to because a new value and application are introduced into the interface, then the collaboration template must be modified and re-compiled, it must be re-deployed, and so forth. A much more flexible implementation stores the pairings of values and destination applications in a database table. To implement this sort of approach, do the following: 1. Create a database table which has one column to store the routing values and another column to store some information that associates the appropriate destination application with the routing value. Table 14 is an example of such a table. Table 14. Routing table example Routing value Customer Destination application value AppA © Copyright IBM Corp. 2001, 2004 145