The Overall Architecture of Oracle


The architecture of Oracle is a combination of background processes, memory structures and datafiles. The Oracle Database is the datafiles which physically store the definitions of database objects and stored data. The Instance containing the memory structures and background processes is used to share the contents of the datafiles amongst many users. The Instance is effectively the database server architecture. The memory structures are used to store the most recently selected data plus the recently used database objects and procedures. This allows for faster access of the most commonly accessed items.

The database structures provide a logical mapping to the datafiles. The data is divided into logical divisions called tablespaces. Some structures are internal to the database and others are external to the database. The internal structure are the processes, the shared memory areas and the database objects. The external structures include the datafiles, the redo and archive logs plus the trace and control file. Datafiles ot tablespaces are often referred to as segments. Segments physically store database objects and data. As already stated a segment is effectively a datafile and is assigned to a logical tablespace within a physical datafile. Segments are divided into extents, physical extensions to segments, and extents are divided into physical blocks. The block is the smallest physical element.

The most significant parts of and facts about an Oracle database are as shown below.

Shown below are some important concepts to note about Oracle.

The architecture of Oracle is as shown below.

The Oracle Database Internal Structure

Oracle is made of internal logical structures and external physical structures. The internal logical structures are generally internal to the Oracle database and perform a function which maps accessibility to the external physical structures.

Capacity Increases in Oracle8

Oracle Parallel Server and Clustered Servers

The benefits of clustered servers are three-fold.

  1. More than one machine is in use and thus greater memory resources are available.
  2. As shown in the diagram below, datafiles are shared by multiple servers. Thus if a server crashes the other(s) servers can still access the datafiles and thus service the users.
  3. Different types of processing can be separated onto the different servers based on critical usage requirements, differing types of usage requirements,etc.
The Architecture of the Parallel Database