JDBC connection pools

A pooling data source uses a collection of persistent connections in order to reduce the overhead of repeatedly opening network connections between the client and server. Opening a new connection for each request is more costly for both the server and the client than keeping a small pool of connections open constantly, ready to be used by new requests. When a request comes in, one of the pre-existing connections in the pool is assigned to it. Only if there are no free connections in the pool is a new connection created. Once the request is complete, the connection returns to the pool and waits to service another request.

The Vertica JDBC driver supports connection pooling as defined in the JDBC 4.0 standard. If you are using a J2EE-based application server in conjunction with Vertica, it should already have a built-in data pooling feature. All that is required is that the application server work with the PooledConnection interface implemented by Vertica's JDBC driver. An application server's pooling feature is usually well-tuned for the works loads that the server is designed to handle. See your application server's documentation for details on how to work with pooled connections. Normally, using pooled connections should be transparent in your code—you will just open connections and the application server will worry about the details of pooling them.

If you are not using an application server, or your application server does not offer connection pooling that is compatible with Vertica, you can use a third-party pooling library, such as the open-source c3p0 or DBCP libraries, to implement connection pooling.