Preliminary understanding of Oracle Architecture

Understanding of Oracle architecture ------- preliminary understanding

First of all, the Oracle architecture has a system diagram, which I also understand. In fact, the architecture is the figure. To understand the architecture, you must understand the figure.

System diagram:

I. Composition of oracle servers

Rdbma is a complete solution, including instance database

Instance: memory structure, background process

Database: Logical Structure of physical files and files

Ii. Memory Structure

1. User Process

The methods for accessing databases are as follows: 1. Access the database directly through a local socket. 2. Establish a connection to the server through TCP three-way handshake, for this connection method, you need to configure $ ORACLE_HOME/NETWORK/ADMIN/TNSNAMES. network Address accessed by ORA 3. remote connection

2. Seerver Process

The client sends a connection to the server to establish a server process. Because the services executed by the user are in the memory, the submitted data should be read from the process to the memory. To achieve this goal, oracle designed the PGA

3. Instance

An instance includes various processes on the server and provides services for the database. An instance only corresponds to one database, but a database can have multiple instances (this is only available in special cases ).

4. PGA (Programe Global Area)

That is, the global area of the program. When a connection is established between the User Process and the database, the oracle database server automatically allocates its memory area, this region is reserved for the session established between the current user and the database, and is a private memory region. It is allocated when the session is created and will be released after the session is closed.

Composition: it includes: UGA (User Global Area) and Stack Space

UGA includes: User Session Date, Cursor Status, and Sort Area

4. SGA (System Global Area)

That is, the global memory area of the system, which is allocated when the instance is started, is mainly used to store the memory area of the database information. The information is shared by the process. As shown in the figure above, the SGA includes: shared Pool, Database Buffer Cache, Java Pool, Redo Log Buffer, Large Pool, and the Shared Pool contains Library Cache and Data Dictionary Cache; the shared pool is mainly used to cache SQL statements executed and data definitions called. The data blocks used by the database are stored in the high-speed buffer zone (that is, the user's high-speed buffer zone), and the high-speed buffer zone is divided into three areas: Dirty data zone (Dirty Buffers) free Buffers and Pinned Buffers)

5. This section describes the process in which a user process submits an SQL query statement to the server and then returns it to the user.

A. the user process first establishes a connection with the server

B. After the connection is established, the user submits the data to the server process, and the server process submits the data information to the UGA in the PGA.

C. The information is transmitted to the instance after being processed by UGA

D. The sharing pool in intance is responsible for processing this statement, and the library cache is responsible for determining whether it is hard analysis or soft analysis.

E. According to the cdo (Optimizer), get the execution plan and prepare the execution statement. The process is: first in the data dictionary cache area, query the table space where the objects in the statement are placed, and the specified row is placed in the block.

F. Start execution. Because this execution needs to be executed in the memory, you need to determine whether the data buffer cache has cached the required block before execution? If yes, the desired result is returned after the data is read in the memory. If not, the server process reads the block from IO to the data buffer cache, then, the ondo block makes the block into an image, and then reads data from the image and returns it to the user.