Connection pooling engine overview

This runtime module provides connection pooling capabilities. With pooling, connections are reused which saves processing time.

Each transport, Socket, HTTP, WebSphere MQ exposes connections that can be pooled this way.

Connections have a set of properties including the target host address but also a host character set and a set of credentials. Two clients are allowed to reuse the same connection if, and only if, they request the exact same set of connection properties.

The Connection pooling engine is written in a transport independent fashion. The actual behavior is driven by an XML configuration file. The legstar-engine-config.xml is an example of such a file. It is a combination of legstar-pooling-config.xml that actually drives the pooling mechanism and the legstar-invoker-config.xml which is shared with the Mainframe Adapter Runtime.

The Connection pooling engine is not a stand-alone application. It has to be started by an initiator. A Servlet initiator is part of the delivery. The Servlet initiator starts when the servlet engine starts and creates an instance of the Engine.

Connection pooling engine is a singleton. It is assumed all clients will run in the same VM.

For instructions on how to install and configure the pooling engine follow this link.