Oracle Database 19c builds on the industry-leading scalability of earlier releases. Oracle's extensive parallel processing is at the heart of its scalability. Not only is parallelism central to data warehousing and query processing, it plays a key role in Oracle's ability to process large volumes of data.
Parallel execution is a commonly used method of speeding up operations by splitting a task into smaller sub tasks. It is key for large scale data processing. Using parallelism, hundreds of terabytes of data can be processed in minutes, not hours or days. Parallel execution uses multiple processes to accomplish a single task. The more effectively the database can leverage all hardware resources—multiple CPUs, multiple IO channels, multiple storage units, multiple nodes in a cluster—the more efficiently queries and other database operations will be processed.
Automatic Degree of Parallelism (Auto DOP) enables the optimizer to automatically decide if a SQL statement should run in parallel and the DOP to use based on the resource requirements of the statement.
Parallel Statement Queuing enables statements to be queued when the required number of processes are not available. Once the required number of processes become available, the SQL statement is dequeued and executed. Parallel Statement Queuing prevents the system from resource saturation and enables optimal utilization of resources.
In-Memory Parallel Execution (IMPX) leverages shared memory (SGA) to store data for subsequent parallel processing. IMPX takes advantage of the ever-increasing memory of today’s database servers; this is especially beneficial on large-scale cluster environments where the aggregated total amount of memory can be multiples of terabytes even when an individual database server "only" holds tens or hundreds of gigabytes of memory.