In a Nutshell

Congratulations! You have just read the longest chapter in this book. We discussed three wait events: latch free, enqueue, and buffer busy waits . They are the in-memory latency.

A latch is a simple locking device used to protect critical SGA memory structures. Every process that reads or modifies data in the SGA must perform the operation under the protection of latches.

An enqueue is a sophisticated locking device used to protect database resources. A resource may be shared if the lock mode is compatible. If not, the requesting process puts its lock request in a queue and it is serviced in order.

A buffer busy waits wait is a wait for a buffer lock or pin. A buffer must be pinned before it can be read or modified so that the integrity of its content can be preserved. A buffer can be pinned by only one process at any one time.