Skip Headers
Oracle® Database Concepts
11g Release 1 (11.1)

Part Number B28318-01
Go to Documentation Home
Go to Book List
Book List
Go to Table of Contents
Go to Index
Go to Master Index
Master Index
Go to Feedback page
Contact Us

Go to previous page
Go to next page
View PDF

9 Process Architecture

This chapter discusses the processes in an Oracle database system and the different configurations available for an Oracle database system.

This chapter contains the following topics:

Introduction to Processes

All connected Oracle Database users must run two modules of code to access an Oracle Database instance.

These code modules are run by processes. A process is a "thread of control" or a mechanism in an operating system that can run a series of steps. (Some operating systems use the terms job or task.) A process normally has its own private memory area in which it runs.

This section includes the following topics:

Multiple-Process Oracle Systems

Multiple-process Oracle (also called multiuser Oracle) uses several processes to run different parts of the Oracle code and additional processes for the users—either one process for each connected user or one or more processes shared by multiple users. Most database systems are multiuser, because one of the primary benefits of a database is managing data needed by multiple users at the same time.

Each process in an Oracle Database instance performs a specific job. By dividing the work of Oracle Database and database applications into several processes, multiple users and applications can connect to a single database instance simultaneously while the system maintains excellent performance.

Types of Processes

The processes in an Oracle Database system can be categorized into two major groups:

  • User processes run the application or Oracle tool code.

  • Oracle Database processes run the Oracle database server code. They include server processes and background processes.

The process structure varies for different Oracle Database configurations, depending on the operating system and the choice of Oracle Database options. The code for connected users can be configured as a dedicated server or a shared server.

With dedicated server, for each user, the database application is run by a different process (a user process) than the one that runs the Oracle database server code (a dedicated server process).

With shared server, the database application is run by a different process (a user process) than the one that runs the Oracle database server code. Each server process that runs Oracle database server code (a shared server process) can serve multiple user processes.

Figure 9-1 illustrates a dedicated server configuration. Each connected user has a separate user process, and several background processes run Oracle Database.

Figure 9-1 An Oracle Database Instance

Description of Figure 9-1 follows
Description of "Figure 9-1 An Oracle Database Instance "

Figure 9-1 can represent multiple concurrent users running an application on the same computer as Oracle Database. This particular configuration usually runs on a mainframe or minicomputer.

See Also:

Overview of User Processes

When a user runs an application program (such as a Pro*C program) or an Oracle tool (such as Oracle Enterprise Manager or SQL*Plus), Oracle Database creates a user process to run the user's application.

Connections and Sessions

Connection and session are closely related to user process but are very different in meaning.

A connection is a communication pathway between a user process and an Oracle Database instance. A communication pathway is established using available interprocess communication mechanisms (on a computer that runs both the user process and Oracle Database) or network software (when different computers run the database application and Oracle Database, and communicate through a network).

A session is a specific connection of a user to an Oracle Database instance through a user process. For example, when a user starts SQL*Plus, the user must provide a valid user name and password, and then a session is established for that user. A session lasts from the time the user connects until the time the user disconnects or exits the database application.

Multiple sessions can be created and exist concurrently for a single Oracle Database user using the same user name. For example, a user with the user name/password of SCOTT/TIGER can connect to the same Oracle Database instance several times.

In configurations without the shared server, Oracle Database creates a server process on behalf of each user session. However, with the shared server, many user sessions can share a single server process.

Overview of Oracle Database Processes

This section describes the two types of processes that run the Oracle database server code (server processes and background processes). It also describes the trace files and alert logs, which record database events for the Oracle Database processes.

This section includes the following topics:

Oracle Database Server Processes

Oracle Database creates server processes to handle the requests of user processes connected to the instance. In some situations when the application and Oracle Database operate on the same computer, it is possible to combine the user process and corresponding server process into a single process to reduce system overhead. However, when the application and Oracle Database operate on different computers, a user process always communicates with Oracle Database through a separate server process.

Server processes (or the server portion of combined user/server processes) created on behalf of each user's application can perform one or more of the following:

  • Parse and run SQL statements issued through the application

  • Read necessary data blocks from datafiles on disk into the shared database buffers of the SGA, if the blocks are not already present in the SGA

  • Return results in such a way that the application can process the information

Oracle Database Background Processes

To maximize performance and accommodate many users, a multiprocess Oracle Database system uses some additional Oracle Database processes called background processes.

An Oracle Database instance can have many background processes; not all are always present. There are numerous background processes. See the V$BGPROCESS view for more information on the background processes. The background processes in an Oracle Database instance can include the following:

On many operating systems, background processes are created automatically when an instance is started.

Figure 9-2 illustrates how each background process interacts with the different parts of an Oracle database, and the rest of this section describes each process.

See Also:

Figure 9-2 Background Processes of a Multiple-Process Oracle Database Instance

Description of Figure 9-2 follows
Description of "Figure 9-2 Background Processes of a Multiple-Process Oracle Database Instance"

Archiver Processes (ARCn)

The archiver processes (ARCn) copy redo log files to a designated storage device after a log switch has occurred. In addition, they can collect transaction redo data and transmit that data to standby destinations. ARCn processes are present only when the database is in ARCHIVELOG mode, and automatic archiving is enabled.

If you anticipate a heavy workload for archiving, such as during bulk loading of data, you can increase the maximum number of archiver processes with the LOG_ARCHIVE_MAX_PROCESSES initialization parameter. The ALTER SYSTEM statement can change the value of this parameter dynamically to increase or decrease the number of ARCn processes.

See Also:

Checkpoint Process (CKPT)

When a checkpoint occurs, Oracle Database must update the headers of all datafiles to record the details of the checkpoint. This is done by the CKPT process. The CKPT process does not write blocks to disk; DBWn always performs that work.

The statistic DBWR checkpoints displayed by the System_Statistics monitor in Oracle Enterprise Manager indicates the number of checkpoint requests completed.

See Also:

Oracle Real Application Clusters Administration and Deployment Guide for information about CKPT with Oracle Real Application Clustersc

Database Writer Process (DBWn)

The database writer process (DBWn) writes the contents of buffers to datafiles. The DBWn processes are responsible for writing modified (dirty) buffers in the database buffer cache to disk. Although one database writer process (DBW0) is adequate for most systems, you can configure additional processes (DBW1 through DBW9 and DBWa through DBWj) to improve write performance if your system modifies data heavily. These additional DBWn processes are not useful on uniprocessor systems.

When a buffer in the database buffer cache is modified, it is marked dirty. A cold buffer is a buffer that has not been recently used according to the least recently used (LRU) algorithm. The DBWn process writes cold, dirty buffers to disk so that user processes are able to find cold, clean buffers that can be used to read new blocks into the cache. As buffers are dirtied by user processes, the number of free buffers diminishes. If the number of free buffers drops too low, user processes that must read blocks from disk into the cache are not able to find free buffers. DBWn manages the buffer cache so that user processes can always find free buffers.

By writing cold, dirty buffers to disk, DBWn improves the performance of finding free buffers while keeping recently used buffers resident in memory. For example, blocks that are part of frequently accessed small tables or indexes are kept in the cache so that they do not need to be read in again from disk. The LRU algorithm keeps more frequently accessed blocks in the buffer cache so that when a buffer is written to disk, it is unlikely to contain data that will be useful soon.

The initialization parameter DB_WRITER_PROCESSES specifies the number of DBWn processes. The maximum number of DBWn processes is 20. If it is not specified by the user during startup, Oracle Database determines how to set DB_WRITER_PROCESSES based on the number of CPUs and processor groups.

The DBWn process writes dirty buffers to disk under the following conditions:

  • When a server process cannot find a clean reusable buffer after scanning a threshold number of buffers, it signals DBWn to write. DBWn writes dirty buffers to disk asynchronously while performing other processing.

  • DBWn periodically writes buffers to advance the checkpoint, which is the position in the redo thread (log) from which instance recovery begins. This log position is determined by the oldest dirty buffer in the buffer cache.

In all cases, DBWn performs batched (multiblock) writes to improve efficiency. The number of blocks written in a multiblock write varies by operating system.

See Also:

Job Queue Processes

Job queue processes are used for batch processing. They run user jobs. They can be viewed as a scheduler service that can be used to schedule jobs as PL/SQL statements or procedures on an Oracle Database instance. Given a start date and an interval, the job queue processes try to run the job at the next occurrence of the interval.

Job queue processes are managed dynamically. This allows job queue clients to use more job queue processes when required. The resources used by the new processes are released when they are idle.

Dynamic job queue processes can run a large number of jobs concurrently at a given interval. The job queue processes run user jobs as they are assigned by the CJQ process. Here's what happens:

  1. The coordinator process, named CJQ0, periodically selects jobs that need to be run from the system JOB$ table. New jobs selected are ordered by time.

  2. The CJQ0 process dynamically spawns job queue slave processes (J000…J999) to run the jobs.

  3. The job queue process runs one of the jobs that was selected by the CJQ process for execution. The processes run one job at a time.

  4. After the process finishes execution of a single job, it polls for more jobs. If no jobs are scheduled for execution, then it enters a sleep state, from which it wakes up at periodic intervals and polls for more jobs. If the process does not find any new jobs, then it aborts after a preset interval.

The initialization parameter JOB_QUEUE_PROCESSES represents the maximum number of job queue processes that can concurrently run on an instance. However, clients should not assume that all job queue processes are available for job execution.


The coordinator process is not started if the initialization parameter JOB_QUEUE_PROCESSES is set to 0.

See Also:

Oracle Database Administrator's Guide for more information about job queues

Log Writer Process (LGWR)

The log writer process (LGWR) is responsible for redo log buffer management—writing the redo log buffer to a redo log file on disk. LGWR writes all redo entries that have been copied into the buffer since the last time it wrote.

The redo log buffer is a circular buffer. When LGWR writes redo entries from the redo log buffer to a redo log file, server processes can then copy new entries over the entries in the redo log buffer that have been written to disk. LGWR normally writes fast enough to ensure that space is always available in the buffer for new entries, even when access to the redo log is heavy.

LGWR writes one contiguous portion of the buffer to disk. LGWR writes:

  • A commit record when a user process commits a transaction

  • Redo log buffers

    • Every three seconds

    • When the redo log buffer is one-third full

    • When a DBWn process writes modified buffers to disk, if necessary


Before DBWn can write a modified buffer, all redo records associated with the changes to the buffer must be written to disk (the write-ahead protocol). If DBWn finds that some redo records have not been written, it signals LGWR to write the redo records to disk and waits for LGWR to complete writing the redo log buffer before it can write out the data buffers.

LGWR writes synchronously to the active mirrored group of redo log files. If one of the files in the group is damaged or unavailable, LGWR continues writing to other files in the group and logs an error in the LGWR trace file and in the system alert log. If all files in a group are damaged, or the group is unavailable because it has not been archived, LGWR cannot continue to function.

When a user issues a COMMIT statement, LGWR puts a commit record in the redo log buffer and writes it to disk immediately, along with the transaction's redo entries. The corresponding changes to data blocks are deferred until it is more efficient to write them. This is called a fast commit mechanism. The atomic write of the redo entry containing the transaction's commit record is the single event that determines the transaction has committed. Oracle Database returns a success code to the committing transaction, although the data buffers have not yet been written to disk.


Sometimes, if more buffer space is needed, LGWR writes redo log entries before a transaction is committed. These entries become permanent only if the transaction is later committed.

When a user commits a transaction, the transaction is assigned a system change number (SCN), which Oracle Database records along with the transaction's redo entries in the redo log. SCNs are recorded in the redo log so that recovery operations can be synchronized in Real Application Clusters and distributed databases.

In times of high activity, LGWR can write to the redo log file using group commits. For example, assume that a user commits a transaction. LGWR must write the transaction's redo entries to disk, and as this happens, other users issue COMMIT statements. However, LGWR cannot write to the redo log file to commit these transactions until it has completed its previous write operation. After the first transaction's entries are written to the redo log file, the entire list of redo entries of waiting transactions (not yet committed) can be written to disk in one operation, requiring less I/O than do transaction entries handled individually. Therefore, Oracle Database minimizes disk I/O and maximizes performance of LGWR. If requests to commit continue at a high rate, then every write (by LGWR) from the redo log buffer can contain multiple commit records.

See Also:

Process Monitor Process (PMON)

The process monitor (PMON) performs process recovery when a user process fails. PMON is responsible for cleaning up the database buffer cache and freeing resources that the user process was using. For example, it resets the status of the active transaction table, releases locks, and removes the process ID from the list of active processes.

PMON periodically checks the status of dispatcher and server processes, and restarts any that have stopped running (but not any that Oracle Database has terminated intentionally). PMON also registers information about the instance and dispatcher processes with the network listener.

Like SMON, PMON checks regularly to see whether it is needed and can be called if another process detects the need for it.

Queue Monitor Processes (QMNn)

The queue monitor process is an optional background process for Oracle Streams Advanced Queuing, which monitors the message queues. You can configure up to 10 queue monitor processes. These processes, like the job queue processes, are different from other Oracle Database background processes in that process failure does not cause the instance to fail.

Recoverer Process (RECO)

The recoverer process (RECO) is a background process used with the distributed database configuration that automatically resolves failures involving distributed transactions. The RECO process of a node automatically connects to other databases involved in an in-doubt distributed transaction. When the RECO process reestablishes a connection between involved database servers, it automatically resolves all in-doubt transactions, removing from each database's pending transaction table any rows that correspond to the resolved in-doubt transactions.

If the RECO process fails to connect with a remote server, RECO automatically tries to connect again after a timed interval. However, RECO waits an increasing amount of time (growing exponentially) before it attempts another connection. The RECO process is present only if the instance permits distributed transactions. The number of concurrent distributed transactions is not limited.

See Also:

Oracle Database Administrator's Guide for more information about distributed transaction recovery

System Monitor Process (SMON)

The system monitor process (SMON) performs recovery, if necessary, at instance startup. SMON is also responsible for cleaning up temporary segments that are no longer in use and for coalescing contiguous free extents within dictionary managed tablespaces. If any terminated transactions were skipped during instance recovery because of file-read or offline errors, SMON recovers them when the tablespace or file is brought back online. SMON checks regularly to see whether it is needed. Other processes can call SMON if they detect a need for it.

With Real Application Clusters, the SMON process of one instance can perform instance recovery for a failed CPU or instance.

Other Oracle Database Background Processes

There are several other background processes that might be running. These can include the following:

  • ACMS (atomic controlfile to memory service) per-instance process is an agent that contributes to ensuring a distributed SGA memory update is either globally committed on success or globally aborted in the event of a failure in an Oracle RAC environment.

  • ASMB is present in a database instance using an Automatic Storage Management disk group. It communicates with the Automatic Storage Management instance.

  • DBRM (database resource manager) process is responsible for setting resource plans and other resource manager related tasks.

    See Also:

    "Overview of the Database Resource Manager" for more information about the database resource manager
  • DIA0 (diagnosability process 0) (only 0 is currently being used) is responsible for hang detection and deadlock resolution.

  • DIAG (diagnosability) process performs diagnostic dumps and executes global oradebug commands.

  • DSKM (slave diskmon) is used as the conduit between RDBMS and ASM instances and the Master Diskmon daemon to communicate I/O Fencing information, I/O Resource Manager Plans, and Transaction Commit Cache information to SAGE storage. It is also used to implement the skgxp ANT protocol between the host nodes and SAGE storage servers. If no SAGE storage is used, the slave diskmon process will exit silently after startup of the instance.

  • EMNC (event monitor coordinator) is the background server process used for database event management and notifications.

  • FBDA (flashback data archiver process) archives the historical rows of tracked tables into flashback data archives. Tracked tables are tables which are enabled for flashback archive. When a transaction containing DML on a tracked table commits, this process stores the pre-image of the rows into the flashback archive. It also keeps metadata on the current rows.

    FBDA is also responsible for automatically managing the flashback data archive for space, organization, and retention and keeps track of how far the archiving of tracked transactions has occurred.

  • GMON maintains disk membership in ASM disk groups.

  • GTX0-j (global transaction) processes provide transparent support for XA global transactions in an Oracle RAC environment. The database autotunes the number of these processes based on the workload of XA global transactions. Global transaction processes are only seen in an Oracle RAC environment.

  • KATE performs proxy I/O to an ASM metafile when a disk goes offline.

  • MARK marks ASM allocation units as stale following a missed write to an offline disk.

  • MMAN is used for internal database tasks.

  • MMNL performs frequent and light-weight manageability-related tasks, such as session history capture and metrics computation.

  • MMON performs various manageability-related background tasks, for example:

    • Issuing alerts whenever a given metrics violates its threshold value

    • Taking snapshots by spawning additional process (MMON slaves)

    • Capturing statistics value for SQL objects which have been recently modified

  • ORBn performs the actual rebalance data extent movements in an Automatic Storage Management instance. There can be many of these at a time, called ORB0, ORB1, and so forth.

  • PSP0 (process spawner) spawns Oracle processes.

  • RBAL coordinates rebalance activity for disk groups in an Automatic Storage Management instance. It performs a global open on Automatic Storage Management disks.

  • SMCO (space management coordinator) process coordinates the execution of various space management related tasks, such as proactive space allocation and space reclamation. It dynamically spawns slave processes (Wnnn) to implement the task.

  • VKTM (virtual keeper of time) is responsible for providing a wall-clock time (updated every second) and reference-time counter (updated every 20 ms and available only when running at elevated priority).

See Also:

Oracle Database Trace Files and the Alert Log

Beginning with Oracle Database 11g, an advanced fault diagnosability infrastructure is included for preventing, detecting, diagnosing, and resolving problems. The problems that are targeted in particular are critical errors such as those caused by database code bugs, metadata corruption, and customer data corruption.

When a critical error occurs, it is assigned an incident number, and diagnostic data for the error (such as trace files) are immediately captured and tagged with this number. The data is then stored in the Automatic Diagnostic Repository (ADR)—a file based repository outside the database—where it can later be retrieved by incident number and analyzed.

Each server and background process can write to an associated trace file. When a process detects an internal error, it dumps information about the error to its trace file. If an internal error occurs and information is written to a trace file, the administrator should contact Oracle Support Services.

All filenames of trace files associated with a background process contain the name of the process that generated the trace file. The one exception to this is trace files generated by job queue processes (Jnnn).

Additional information in trace files can provide guidance for tuning applications or an instance. Background processes always write this information to a trace file when appropriate.

Each database also has an alert.log. The alert log of a database is a chronological log of messages and errors, including the following:

  • All internal errors (ORA-600), block corruption errors (ORA-1578), and deadlock errors (ORA-60) that occur

  • Administrative operations, such as the SQL statements CREATE/ALTER/DROP DATABASE/TABLESPACE and the Oracle Enterprise Manager or SQL*Plus statements STARTUP, SHUTDOWN, ARCHIVE LOG, and RECOVER

  • Several messages and errors relating to the functions of shared server and dispatcher processes

  • Errors during the automatic refresh of a materialized view

Oracle Database uses the alert log to keep a record of these events as an alternative to displaying the information on an operator's console. (Many systems also display this information on the console.) If an administrative operation is successful, a message is written in the alert log as "completed" along with a time stamp.

See Also:

Shared Server Architecture

Shared server architecture eliminates the need for a dedicated server process for each connection. A dispatcher directs multiple incoming network session requests to a pool of shared server processes. An idle shared server process from a shared pool of server processes picks up a request from a common queue, which means a small number of shared servers can perform the same amount of processing as many dedicated servers. Also, because the amount of memory required for each user is relatively small, less memory and process management are required, and more users can be supported.

A number of different processes are needed in a shared server system:

Shared server processes require Oracle Net Services or SQL*Net version 2.


To use shared servers, a user process must connect through Oracle Net Services or SQL*Net version 2, even if the process runs on the same computer as the Oracle Database instance.

When an instance starts, the network listener process opens and establishes a communication pathway through which users connect to Oracle Database. Then, each dispatcher process gives the listener process an address at which the dispatcher listens for connection requests. At least one dispatcher process must be configured and started for each network protocol that the database clients will use.

When a user process makes a connection request, the listener examines the request and determines whether the user process can use a shared server process. If so, the listener returns the address of the dispatcher process that has the lightest load, and the user process connects to the dispatcher directly.

Some user processes cannot communicate with the dispatcher, so the network listener process cannot connect them to a dispatcher. In this case, or if the user process requests a dedicated server, the listener creates a dedicated server and establishes an appropriate connection.

The Oracle Database shared server architecture increases the scalability of applications and the number of clients simultaneously connected to the database. It can enable existing applications to scale up without making any changes to the application itself.

This section includes the following topics:

Dispatcher Request and Response Queues

A request from a user is a single program interface call that is part of the user's SQL statement. When a user makes a call, its dispatcher places the request on the request queue, where it is picked up by the next available shared server process.

The request queue is in the SGA and is common to all dispatcher processes of an instance. The shared server processes check the common request queue for new requests, picking up new requests on a first-in-first-out basis. One shared server process picks up one request in the queue and makes all necessary calls to the database to complete that request.

When the server completes the request, it places the response on the calling dispatcher's response queue. Each dispatcher has its own response queue in the SGA. The dispatcher then returns the completed request to the appropriate user process.

For example, in an order entry system each clerk's user process connects to a dispatcher and each request made by the clerk is sent to that dispatcher, which places the request in the request queue. The next available shared server process picks up the request, services it, and puts the response in the response queue. When a clerk's request is completed, the clerk remains connected to the dispatcher, but the shared server process that processed the request is released and available for other requests. While one clerk is talking to a customer, another clerk can use the same shared server process.

Figure 9-3 illustrates how user processes communicate with the dispatcher across the program interface and how the dispatcher communicates users' requests to shared server processes.

Figure 9-3 The Shared Server Configuration and Processes

Description of Figure 9-3 follows
Description of "Figure 9-3 The Shared Server Configuration and Processes"

This section includes the following topics:

Dispatcher Processes (Dnnn)

The dispatcher processes support shared server configuration by allowing user processes to share a limited number of server processes. With the shared server, fewer shared server processes are required for the same number of users, Therefore, the shared server can support a greater number of users, particularly in client/server environments where the client application and server operate on different computers.

You can create multiple dispatcher processes for a single database instance. At least one dispatcher must be created for each network protocol used with Oracle Database. The database administrator starts an optimal number of dispatcher processes depending on the operating system limitation and the number of connections for each process, and can add and remove dispatcher processes while the instance runs.


Each user process that connects to a dispatcher must do so through Oracle Net Services or SQL*Net version 2, even if both processes are running on the same computer.

In a shared server configuration, a network listener process waits for connection requests from client applications and routes each to a dispatcher process. If it cannot connect a client application to a dispatcher, the listener process starts a dedicated server process, and connects the client application to the dedicated server. The listener process is not part of an Oracle Database instance; rather, it is part of the networking processes that work with Oracle Database.

See Also:

Shared Server Processes (Snnn)

Each shared server process serves multiple client requests in the shared server configuration. Shared server processes and dedicated server processes provide the same functionality, except shared server processes are not associated with a specific user process. Instead, a shared server process serves any client request in the shared server configuration.

The PGA of a shared server process does not contain user-related data (which needs to be accessible to all shared server processes). The PGA of a shared server process contains only stack space and process-specific variables.

All session-related information is contained in the SGA. Each shared server process needs to be able to access all sessions' data spaces so that any server can handle requests from any session. Space is allocated in the SGA for each session's data space. You can limit the amount of space that a session can allocate by setting the resource limit PRIVATE_SGA to the desired amount of space in the user's profile.

Oracle Database dynamically adjusts the number of shared server processes based on the length of the request queue. The number of shared server processes that can be created ranges between the values of the initialization parameters SHARED_SERVERS and MAX_SHARED_SERVERS.

See Also:

Restricted Operations of the Shared Server

Certain administrative activities cannot be performed while connected to a dispatcher process, including shutting down or starting an instance and media recovery. An error message is issued if you attempt to perform these activities while connected to a dispatcher process.

These activities are typically performed when connected with administrator privileges. When you want to connect with administrator privileges in a system configured with shared servers, you must state in your connect string that you want to use a dedicated server process (SERVER=DEDICATED) instead of a dispatcher process.

See Also:

Dedicated Server Configuration

Figure 9-4 illustrates Oracle Database running on two computers using the dedicated server architecture. In this configuration, a user process runs the database application on one computer, and a server process runs the associated Oracle database server on another computer.

Figure 9-4 Oracle Database Using Dedicated Server Processes

Description of Figure 9-4 follows
Description of "Figure 9-4 Oracle Database Using Dedicated Server Processes"

The user and server processes are separate, distinct processes. The separate server process created on behalf of each user process is called a dedicated server process (or shadow process), because this server process acts only on behalf of the associated user process.

This configuration maintains a one-to-one ratio between the number of user processes and server processes. Even when the user is not actively making a database request, the dedicated server process remains (though it is inactive and can be paged out on some operating systems).

Figure 9-4 shows user and server processes running on separate computers connected across a network. However, the dedicated server architecture is also used if the same computer runs both the client application and the Oracle database server code but the host operating system could not maintain the separation of the two programs if they were run in a single process. UNIX is a common example of such an operating system.

In the dedicated server configuration, the user and server processes communicate using different mechanisms:

See Also:

for more information about communication links

Database Resident Connection Pooling

Database Resident Connection Pooling (DRCP) provides a connection pool in the database server for typical Web application usage scenarios. DRCP pools dedicated servers, which comprise of a server foreground combined with a database session, to create pooled servers.

A Web application typically acquires a database connection, uses the connection for a short period, and then releases the connection. DRCP enables multiple Web application threads and processes to share the pooled servers for their connection needs.

DRCP complements middle-tier connection pools that share connections between threads in a middle-tier process. DRCP also enables you to share database connections across multiple middle-tier processes. These middle-tier processes may belong to the same or different middle-tier host.

DRCP enables a significant reduction in key database resources that are required to support a large number of client connections. DRCP reduces the amount of memory required for the database server and boosts the scalability of both the database server and the middle-tier. The pool of readily available servers also reduces the cost of re-creating client connections.

DRCP is especially useful for architectures with multi-process, single-threaded application servers, such as PHP and Apache servers, that cannot do middle-tier connection pooling. The database can scale to tens of thousands of simultaneous connections with DRCP.

The pooled server model closely follows the dedicated model, which is used to connect to Oracle by default. The pooled server model does away with the overhead of dedicating a server for every connection that requires the server for a short period. DRCP allows a connection to acquire, and then voluntarily release the pooled server for use by other similar connections. On being acquired by a connection, a pooled server essentially transforms into a dedicated server for that connection until it is released back into the pool. Clients getting connections out of the database resident connection pool connect to an Oracle background process known as the connection broker. The connection broker implements the pool functionality and multiplexes pooled servers among inbound connections from client processes.

When a client needs to perform some backend database work, the connection broker picks up a pooled server from the pool and assigns it to the client. Subsequently, the client is directly connected to the pooled server until the request is served. After the server has finished processing the client request, the server goes back into the pool and the connection from the client is restored to the connection broker process. Figure 9-5 illustrates database resident connection pooling.

Figure 9-5 Pool of Dedicated Server Processes Handling Connections Through the Connection Broker Process

Description of Figure 9-5 follows
Description of "Figure 9-5 Pool of Dedicated Server Processes Handling Connections Through the Connection Broker Process"

Using Database Resident Connection Pooling

Database resident connection pooling enables you to freely scale your middle-tier hardware without worrying about running out of memory on the database side. This is because a smaller pool of dedicated server processes can serve a larger number of middle-tier processes.

The default connection pool is called, SYS_DEFAULT_CONNECTION_POOL. To use database resident connection pooling, the database administrator needs to explicitly start the pool. The following example illustrates this:

Log into SQL*Plus as SYSDBA

Run the following command:



Currently, only the default connection pool is supported.

To connect to the shared pool, the server type should be set to POOLED in the database connection string. For example:


You can also connect to the shared pool using the easy connect naming method. You need to use the keyword POOLED with the database service name. For example:

CONNECT joeuser/

This section includes the following topics:

Connection Classes

A connection class defines a logical name for the type of connection required by the application. Two different users cannot share connections, or sessions, amongst themselves. For example, a session first created for user HR is only given out to subsequent requests on behalf of the user HR. A connection class enables further separation between the sessions of a given user. The connection class lets different applications, connecting as the same database user, identify their sessions using a logical name that corresponds to the application. DRCP then ensures that sessions belonging to a particular connection class are not shared outside the connection class.

Session Purity

Session purity specifies whether the application requires a brand new session (PURITY=NEW), or whether the application logic is set up to reuse a pooled session (PURITY=SELF). If the application can reuse a pooled session, then a free session with the requested connection class is allotted to the application.

Connection classes and session purity are specified by the client as attributes of a DRCP connection. The default connection class value is username.SHARED. By default, sessions with the same username are shared when purity is SELF.

The default value for purity is NEW. The defaults can differ for different application scenarios. Please see the respective application manuals for details on using DRCP with your application.

See Also:

Oracle Call Interface Programmer's Guide for details on using connection classes and session purity

The Program Interface

The program interface is the software layer between a database application and Oracle Database. The program interface:

The Oracle code acts as a server, performing database tasks on behalf of an application (a client), such as fetching rows from data blocks. It consists of several parts, provided by both Oracle Database software and operating system-specific software.

This section includes the following topics:

Program Interface Structure

The program interface consists of the following pieces:

  • Oracle call interface (OCI) or the Oracle runtime library (SQLLIB)

  • The client or user side of the program interface

  • Various Oracle Net Services drivers (protocol-specific communications software)

  • Operating system communications software

  • The server or Oracle Database side of the program interface (also called the OPI)

Both the user and Oracle Database sides of the program interface run Oracle software, as do the drivers.

Oracle Net Services is the portion of the program interface that allows the client application program and the Oracle database server to reside on separate computers in your communication network.

Program Interface Drivers

Drivers are pieces of software that transport data, usually across a network. They perform operations such as connect, disconnect, signal errors, and test for errors. Drivers are specific to a communications protocol, and there is always a default driver.

You can install multiple drivers (such as the asynchronous or DECnet drivers) and select one as the default driver, but allow an individual user to use other drivers by specifying the desired driver at the time of connection. Different processes can use different drivers. A single process can have concurrent connections to a single database or to multiple databases (either local or remote) using different Oracle Net Services drivers.

See Also:

  • Your system installation and configuration guide for details about choosing, installing, and adding drivers

  • Your system Oracle Net Services documentation for information about selecting a driver at runtime while accessing Oracle Database

  • Oracle Database Net Services Administrator's Guide

Communications Software for the Operating System

The lowest-level software connecting the user side to the Oracle Database side of the program interface is the communications software, which is provided by the host operating system. DECnet, TCP/IP, LU6.2, and ASYNC are examples. The communication software can be supplied by Oracle, but it is usually purchased separately from the hardware vendor or a third-party software supplier.

See Also:

Your Oracle Database operating system-specific documentation for more information about the communication software of your system