How does DRCP (Dynamic Response Connection Pooling) work?

DRCP (Dynamic Response Connection Pooling) is an Oracle Database mechanism that allows multiple client processes to share a pool of lightweight server processes (brokers). Instead of allocating a separate heavyweight server process for each client connection, DRCP uses a broker that manages a pool of slave server processes.

When a client connects to DRCP, it gains access to one of the slaves from the pool via the broker. After the transaction is completed or the query is executed, the slave is released back into the pool, becoming available for other clients. The broker manages the lifecycle of slaves, creating and destroying them based on load, which allows efficient use of server resources and reduces overhead associated with creating and maintaining multiple server processes.

Advantages of DRCP:

• Efficient use of server resources.
• Reduction in the number of server processes.
• Improved scalability as the number of client connections grows.
• Reduced overhead for each connection.

DRCP configuration includes setting up the pool, defining maximum and minimum number of slaves, timeouts, and other parameters.

Example configuration:

-- Starting the DRCP pool
EXEC DBMS_CONNECTION_POOL.START_POOL;

-- Setting the minimum number of processes in the pool
EXEC DBMS_CONNECTION_POOL.CONFIGURE_POOL('MIN_SIZE', '10');

-- Setting the maximum number of processes in the pool
EXEC DBMS_CONNECTION_POOL.CONFIGURE_POOL('MAX_SIZE', '100');

-- Retrieving current configuration
SELECT * FROM V$CPOOL_CONFIG;

On the client side, connection to DRCP is specified in the connection string, for example, using (SERVER=POOLED).

DRCP is especially useful in scenarios with a large number of short-lived connections, such as web applications, where frequent connect and disconnect operations can create significant load without DRCP.