connection pooling using DB2 connect or an application server
Help



DB2 Connect Connection Pooling versus Application Server Connection Pooling

Connection pooling is a must for any web technologies based application that is to support large volumes of transactions. Most web application servers now provide their own way of pooling database connections. For example, both Microsoft(R) MTS (COM+) and IBM WebSphere(R) provide connection pooling.

Application pooling mechanisms implemented by these servers differ significantly from what is provided by the DB2 Connect servers. Since application servers pool connections only for their own use they typically presume that user id, password, isolation levels and so on will be exactly the same for all connections. Even more important, application servers only pool connections initiated by the same process. This means that connections from other machines, users or processes are not pooled. While these application server pooling techniques are effective for reusing connections established by the same instance of an application they are absolutely ineffective for pooling connections from multiple users, servers etc.

Connection pooling, provided by the DB2 Connect servers, is completely application, machine and user independent. Connections from multiple clients, application servers all with different user ids can all reuse each other's connections resulting in a much better utilization of the pooled resources.

Which type of connection pooling is the right one to use? Both. Generally, using both DB2 Connect connection pooling and Application Server connection pooling is a good strategy since they don't interfere with each other. Even when application server connection pooling is enabled, DB2 Connect connection pooling can provide connection reuse for multiple application servers as well as other clients using the DB2 Connect server.

- posted by ThinkEddy @ 1:13 PM