Oracle Database Architecture

For allmost all the application, written today, needs a database to keep the information centralized. For this purpose, a database management system may be used which will help to manage data effectively and efficiently with fast response time.

Oracle is one such famous database management system which supports many activities and supports the required efficiency and accuracy from the architecture itself.

Oracle Relational Database Management System (RDBMS) is consists of two major layers which supports for high response rate: physical layer and logical layer.

What actually happens in the physical layer is, it stores the data as it is, without any relationship among data. In simple terms if anyone peeps in to the physical layer he/she might see some data blocks bt without any idea of relationship among them.

Physical layer consists of three different file types, data files, redo log files and control files. Data files contain the information about the database, basically the raw data which resides in the database. There can be one or more datafiles present in the database and having many of them will lead to enhanced performance of the RDBMS. The maximum amount of datafiles which can be created is limited by the oracle from the variable, "MAXDATAFILES".

Second type of files in the physical layer is the "redo log" files. These files contain the data which can be used to recover the database in a situation of a database failure. It is more like a log which has all the changes of the system. And the important fact to remember is, if the redo log files are lost there is no way of recovering the database. Important to make sure the safety of these files.

Control files are the third type of files in the physical layer which has the information needed to start an instance of the oracle database. For an example it has the information of location of the redo log file etc. Since this is important aspect of the database architecture, multiple copies ot the file is saved in different places.

In logical layer of the orcle architecture, it consists of two layers inside, tablespaces and database schema. The main purpose of the tablespace is to logically group the data. In applications wt we do is to access tables in the database. That table relationship is created by this logical layer which will logically integrate the data and show the applications the relationship in data. Using this logical layer make the administration easy. And also it provide the facility for the data blocks, data files to be spread over a wide diskspace rather than having them in the same disk. Having datafiles in seperate disks have a really good advantage on performance factor as it will cause less traffic in I/O.
Some of the important concepts in this tablespace are "data block", "extent" and "segment". Data block is the actual physical size of one data block in oracle database. The contiguous data blocks are called as extent whereas a set of extents is called as segment.

Database schema is the next layer inside logical layer where the oracle instances dominate. Oracle instances are made via the System Global Area (SGA) and it is a combination of background processes and memory buffers.
SGN referes to share information where data plus control information resides in that. Program Global Area (PGA) is another word that comes along with SGA. Other than the fact that SGA is shared information and PGA is about non shared information most of things in these two are similar.