Understanding Oracle Database Architecture: Components, Memory, and Background Processes

Overview

Oracle database servers are divided into two primary parts: the Oracle instance (the upper half) and the database (the lower half). Users do not connect directly to the database; instead a user process communicates with a server process, which interacts with the instance, which finally accesses the underlying database.

The interaction flow is illustrated in the second diagram.

Instance and Memory Structure

Oracle server = instance + database.

Instance = SGA (shared global area) + background processes.

Memory = SGA + PGA.

SGA consists of:

Database buffer cache – stores copies of data blocks read from data files, enabling fast access and dynamic size management.

Redo log buffer – records all changes to database blocks before they are written to redo log files; size is controlled by log_buffer.

Shared pool – includes the library cache (SQL and PL/SQL statements) and the data dictionary cache (metadata such as tables, indexes, users, privileges).

Large pool – optional area used by shared server sessions to reduce pressure on the shared pool; size can be adjusted dynamically.

Java pool – optional memory for Java execution within the database.

Streams pool – optional memory for Oracle Streams and related features.

Background Processes

PMON (process monitor) – cleans up after failed user sessions, rolls back transactions, releases locks, and deregisters dead processes.

SMON (system monitor) – performs instance recovery, rolls forward redo logs, and releases temporary tablespace.

DBWn (database writer) – writes dirty blocks from the buffer cache to data files when a checkpoint occurs, the cache reaches a threshold, or space is needed.

LGWR (log writer) – flushes the redo log buffer to online redo log files on commit, when the buffer is one‑third full, reaches 1 MB, or every three seconds.

CKPT (checkpoint) – signals DBWn to write dirty blocks, updates the SCN in data file headers and control files, and runs roughly every three seconds to maintain consistency.

ARCn (archiver) – copies online redo log files to archive log files when archiving mode is enabled, ensuring that redo logs are available for recovery.