|Oracle® Streams Advanced Queuing User's Guide and Reference
10g Release 2 (10.2)
|PDF · Mobi · ePub|
This chapter discusses Oracle Streams Advanced Queuing (AQ) and the requirements for complex information handling in an integrated environment.
This chapter contains the following topics:
When Web-based business applications communicate with each other, producer applications enqueue messages and consumer applications dequeue messages. At the most basic level of queuing, one producer enqueues one or more messages into one queue. Each message is dequeued and processed once by one of the consumers. A message stays in the queue until a consumer dequeues it or the message expires. A producer may stipulate a delay before the message is available to be consumed, and a time after which the message expires. Likewise, a consumer may wait when trying to dequeue a message if no message is available. An agent program or application may act as both a producer and a consumer.
Producers can enqueue messages in any sequence. Messages are not necessarily dequeued in the order in which they are enqueued. Messages can be enqueued without being dequeued.
At a slightly higher level of complexity, many producers enqueue messages into a queue, all of which are processed by one consumer. Or many producers enqueue messages, each message being processed by a different consumer depending on type and correlation identifier.
Enqueued messages are said to be propagated when they are reproduced on another queue, which can be in the same database or in a remote database.
Applications often use data in different formats. A transformation defines a mapping from one data type to another. The transformation is represented by a SQL function that takes the source data type as input and returns an object of the target data type. You can arrange transformations to occur when a message is enqueued, when it is dequeued, or when it is propagated to a remote subscriber.
Oracle Streams AQ provides database-integrated message queuing functionality. It is built on top of Oracle Streams and leverages the functions of Oracle Database so that messages can be stored persistently, propagated between queues on different computers and databases, and transmitted using Oracle Net Services and HTTP(S).
Because Oracle Streams AQ is implemented in database tables, all operational benefits of high availability, scalability, and reliability are also applicable to queue data. Standard database features such as recovery, restart, and security are supported by Oracle Streams AQ. You can use database development and management tools such as Oracle Enterprise Manager to monitor queues. Like other database tables, queue tables can be imported and exported.
Messages can be queried using standard SQL. This means that you can use SQL to access the message properties, the message history, and the payload. With SQL access you can also do auditing and tracking. All available SQL technology, such as indexes, can be used to optimize access to messages.
Note:Oracle Streams AQ does not support data manipulation language (DML) operations on a queue table or an associated index-organized table (IOT), if any. The only supported means of modifying queue tables is through the supplied APIs. Queue tables and IOTs can become inconsistent and therefore effectively ruined, if DML operations are performed on them.
Oracle Streams AQ supports system-level access control for all queuing operations, allowing an application designer or DBA to designate users as queue administrators. A queue administrator can invoke Oracle Streams AQ administrative and operational interfaces on any queue in the database. This simplifies administrative work because all administrative scripts for the queues in a database can be managed under one schema.
Oracle Streams AQ supports queue-level access control for enqueue and dequeue operations. This feature allows the application designer to protect queues created in one schema from applications running in other schemas. The application designer can grant only minimal access privileges to applications that run outside the queue schema.
Requests for service must be decoupled from supply of services to increase efficiency and enable complex scheduling. Oracle Streams AQ exhibits high performance as measured by:
Number of messages enqueued and dequeued each second
Time to evaluate a complex query on a message warehouse
Time to recover and restart the messaging process after a failure
Queuing systems must be scalable. Oracle Streams AQ exhibits high performance when the number of programs using the application increases, when the number of messages increases, and when the size of the message warehouse increases.
Messages that constitute requests for service must be stored persistently and processed exactly once for deferred execution to work correctly in the presence of network, computer, and application failures. Oracle Streams AQ is able to meet requirements in the following situations:
Applications do not have the resources to handle multiple unprocessed messages arriving simultaneously from external clients or from programs internal to the application.
Communication links between databases are not available all the time or are reserved for other purposes. If the system falls short in its capacity to deal with these messages immediately, then the application must be able to store the messages until they can be processed.
External clients or internal programs are not ready to receive messages that have been processed.
Queuing systems must deal with priorities, and those priorities can change:
Messages arriving later can be of higher priority than messages arriving earlier.
Messages may wait for later messages before actions are taken.
The same message may be accessed by different processes.
Messages in a specific queue can become more important, and so must be processed with less delay or interference from messages in other queues.
Messages sent to some destinations can have a higher priority than others.
Queuing systems must preserve message metadata, which can be as important as the payload data. For example, the time that a message is received or dispatched can be crucial for business and legal reasons. With the persistence features of Oracle Streams AQ, you can analyze periods of greatest demand or evaluate the lag between receiving and completing an order.
Oracle Streams AQ supports enqueue, dequeue, and propagation operations where the queue type is an abstract datatype, ADT. It also supports enqueue and dequeue operations if the types are inherited types of a base ADT. Propagation between two queues where the types are inherited from a base ADT is not supported.
Oracle Streams AQ also supports
ANYDATA queues, which enable applications to enqueue different message types in a single queue.
If you plan to enqueue, propagate, or dequeue user-defined type messages, then each type used in these messages must exist at every database where the message can be enqueued in a queue. Some environments use directed networks to route messages through intermediate databases before they reach their destination. In such environments, the type must exist at each intermediate database, even if the messages of this type are never enqueued or dequeued at a particular intermediate database.
In addition, the following requirements must be met for such types:
Type name must be the same at each database.
Type must be in the same schema at each database.
Shape of the type must match exactly at each database.
Type cannot use inheritance or type evolution at any database.
Type cannot contain varrays, nested tables, LOBs, rowids, or urowids.
The object identifier need not match at each database.
You can use object types to structure and manage message payloads. Relational database systems in general have a richer typing system than messaging systems. Because Oracle Database is an object-relational database system, it supports traditional relational and user-defined types. Many powerful features are enabled as a result of having strongly typed content, such as content whose format is defined by an external type system. These include:
Oracle Streams AQ can examine the content and automatically route the message to another queue based on the content.
A publish and subscribe system is built on top of a messaging system so that you can create subscriptions based on content.
The ability to run queries on the content of the message enables message warehousing.
You can create queues that use the new opaque type,
XMLType. These queues can be used to transmit and store messages that are XML documents. Using
XMLType, you can do the following:
Store any type of message in a queue
documents internally as CLOB objects
more than one type of payload in a queue
Query XMLType columns using the operator
Specify the operators in subscriber rules or dequeue conditions
You can register system events, user events, and notifications on queues with Oracle Internet Directory. System events are database startup, database shutdown, and system error events. User events include user log on and user log off, DDL statements (create, drop, alter), and DML statement triggers. Notifications on queues include OCI notifications, PL/SQL notifications, and e-mail notifications.
You can also create aliases for Oracle Streams AQ agents in Oracle Internet Directory. These aliases can be specified while performing Oracle Streams AQ enqueue, dequeue, and notification operations. This is useful when you do not want to expose an internal agent name.
Real Application Clusters can be used to improve Oracle Streams AQ performance by allowing different queues to be managed by different instances. You do this by specifying different instance affinities (preferences) for the queue tables that store the queues. This allows queue operations (enqueue and dequeue) on different queues to occur in parallel.
If compatibility is set to Oracle8i release 8.1.5 or higher, then an application can specify the instance affinity for a queue table. When Oracle Streams AQ is used with Real Application Clusters and multiple instances, this information is used to partition the queue tables between instances for queue-monitor scheduling as well as for propagation. The queue table is monitored by the queue monitors of the instance specified by the user. If the owner of the queue table is terminated, then the secondary instance or some available instance takes over the ownership for the queue table.
If an instance affinity is not specified, then the queue tables are arbitrarily partitioned among the available instances. This can result in pinging between the application accessing the queue table and the queue monitor monitoring it. Specifying the instance affinity prevents this, but does not prevent the application from accessing the queue table and its queues from other instances.
Oracle Streams AQ provides the message management and communication needed for application integration. In an integrated environment, messages travel between the Oracle Database server, applications, and users, as shown in Figure 1-1.
Messages are exchanged between a client and the Oracle Database server or between two Oracle Database servers using Oracle Net Services. Oracle Net Services also propagates messages from one Oracle Database queue to another. Or, as shown in Figure 1-1, you can perform Oracle Streams AQ operations over the Internet using HTTP(S). In this case, the client, a user or Internet application, produces structured XML messages. During propagation over the Internet, Oracle Database servers communicate using structured XML also.
Application integration also involves the integration of heterogeneous messaging systems. Oracle Streams AQ seamlessly integrates with existing non-Oracle Database messaging systems like IBM WebSphere MQ through Messaging Gateway, thus allowing existing WebSphere MQ-based applications to be integrated into an Oracle Streams AQ environment.
This section contains these topics:
Client/Server applications usually run in a synchronous manner. Figure 1-2 demonstrates the asynchronous alternative using Oracle Streams AQ. In this example Application B (a server) provides service to Application A (a client) using a request/response queue.
Application A enqueues a request into the request queue. In a different transaction, Application B dequeues and processes the request. Application B enqueues the result in the response queue, and in yet another transaction, Application A dequeues it.
The client need not wait to establish a connection with the server, and the server dequeues the message at its own pace. When the server is finished processing the message, there is no need for the client to be waiting to receive the result. A process of double-deferral frees both client and server.
A message can only be enqueued into one queue at a time. If a producer had to insert the same message into several queues in order to reach different consumers, then this would require management of a very large number of queues. To allow multiple consumers to dequeue the same message, Oracle Streams AQ provides for queue subscribers and message recipients.
To allow for subscriber and recipient lists, the queue must reside in a queue table that is created with the multiple consumer option. Each message remains in the queue until it is consumed by all its intended consumers.
Multiple consumers, which can be either applications or other queues, can be associated with a queue as subscribers. This causes all messages enqueued in the queue to be made available to be consumed by each of the queue subscribers. The subscribers to the queue can be changed dynamically without any change to the messages or message producers.
You cannot add subscriptions to single-consumer queues or exception queues. A consumer that is added as a subscriber to a queue is only able to dequeue messages that are enqueued after the subscriber is added. No two subscribers can have the same values for name, address, and protocol. At least one of these attributes must be different for two subscribers.
It cannot be known which subscriber will dequeue which message first, second, and so on, because there is no priority among subscribers. More formally, the order of dequeuing by subscribers is indeterminate.
Subscribers can also be rule-based. Similar in syntax to the
WHERE clause of a SQL query, rules are expressed in terms of attributes that represent message properties or message content. These subscriber rules are evaluated against incoming messages, and those rules that match are used to determine message recipients.
In Figure 1-3, Application B and Application C each need messages produced by Application A, so a multiconsumer queue is specially configured with Application B and Application C as queue subscribers. Each receives every message placed in the queue.
A message producer can submit a list of recipients at the time a message is enqueued. This allows for a unique set of recipients for each message in the queue. The recipient list associated with the message overrides the subscriber list associated with the queue, if there is one. The recipients need not be in the subscriber list. However, recipients can be selected from among the subscribers.
A recipient can be specified only by its name, in which case the recipient must dequeue the message from the queue in which the message was enqueued. It can be specified by its name and an address with a protocol value of 0. The address should be the name of another queue in the same database or another installation of Oracle Database (identified by the database link), in which case the message is propagated to the specified queue and can be dequeued by a consumer with the specified name. If the recipient's name is NULL, then the message is propagated to the specified queue in the address and can be dequeued by the subscribers of the queue specified in the address. If the protocol field is nonzero, then the name and address are not interpreted by the system and the message can be dequeued by a special consumer.
Subscribing to a queue is like subscribing to a magazine: each subscriber is able to dequeue all the messages placed into a specific queue, just as each magazine subscriber has access to all its articles. Being a recipient, on the other hand, is like getting a letter: each recipient is a designated target of a particular message.
Figure 1-4 shows how Oracle Streams AQ can accommodate both kinds of consumers. Application A enqueues messages. Application B and Application C are subscribers. But messages can also be explicitly directed toward recipients like Application D, which may or may not be subscribers to the queue. The list of such recipients for a given message is specified in the enqueue call for that message. It overrides the list of subscribers for that queue.
Note:Multiple producers can simultaneously enqueue messages aimed at different targeted recipients.
Application A begins a workflow by enqueuing Message 1.
Application B dequeues it, performs whatever activity is required, and enqueues Message 2.
Application C dequeues Message 2 and generates Message 3.
Application D, the final step in the workflow, dequeues it.
Note:The contents of the messages 1, 2 and 3 can be the same or different. Even when they are different, messages can contain parts of the contents of previous messages.
The queues are used to buffer the flow of information between different processing stages of the business process. By specifying delay interval and expiration time for a message, a window of execution can be provided for each of the applications.
From a workflow perspective, knowledge of the volume and timing of message flows is a business asset quite apart from the value of the payload data. Oracle Streams AQ helps you gain this knowledge by supporting the optional retention of messages for analysis of historical patterns and prediction of future trends.
A point-to-point message is aimed at a specific target. Senders and receivers decide on a common queue in which to exchange messages. Each message is consumed by only one receiver. Figure 1-6 shows that each application has its own message queue, known as a single-consumer queue.
A publish/subscribe message can be consumed by multiple receivers, as shown in Figure 1-7. Publish/subscribe messaging has a wide dissemination mode called broadcast and a more narrowly aimed mode called multicast.
Broadcasting is like a radio station not knowing exactly who the audience is for a given program. The dequeuers are subscribers to multiconsumer queues. In contrast, multicast is like a magazine publisher who knows who the subscribers are. Multicast is also referred to as point-to-multipoint, because a single publisher sends messages to multiple receivers, called recipients, who may or may not be subscribers to the queues that serve as exchange mechanisms.
Publish/subscribe describes a situation in which a publisher application enqueues messages to a queue anonymously (no recipients specified). The messages are then delivered to subscriber applications based on rules specified by each application. The rules can be defined on message properties, message data content, or both.
Set up one or more queues to hold messages. These queues should represent an area or subject of interest. For example, a queue can be used to represent billed orders.
Set up a set of rule-based subscribers. Each subscriber can specify a rule which represents a specification for the messages that the subscriber wishes to receive. A null rule indicates that the subscriber wishes to receive all messages.
Publisher applications publish messages to the queue by invoking an enqueue call.
Subscriber applications can receive messages with a dequeue call. This retrieves messages that match the subscription criteria.
Subscriber applications can also use a listen call to monitor multiple queues for subscriptions on different queues. This is a more scalable solution in cases where a subscriber application has subscribed to many queues and wishes to receive messages that arrive in any of the queues.
Subscriber applications can also use the Oracle Call Interface (OCI) notification mechanism. This allows a push mode of message delivery. The subscriber application registers the queues (and subscriptions specified as subscribing agent) from which to receive messages. This registers a callback to be invoked when messages matching the subscriptions arrive.
Figure 1-8 illustrates the use of Oracle Streams AQ for implementing a publish/subscribe relationship between publisher Application A and subscriber Applications B, C, and D:
Application B subscribes with rule "priority = 1".
Application C subscribes with rule "priority > 1".
Application D subscribes with rule "priority = 3".
If Application A enqueues three messages with priorities 1, 2, and 3 respectively, then the messages will be delivered as follows:
Application B receives a single message (priority 1).
Application C receives two messages (priority 2, 3).
Application D receives a single message (priority 3).
Buffered messaging, a new feature in Oracle Streams AQ 10g Release 2 (10.2), combines the rich functionality that this product has always offered with a much faster queuing implementation. Buffered messaging is ideal for applications that do not require the reliability and transaction support of Oracle Streams AQ persistent messaging.
Buffered messaging is faster than persistent messaging, because its messages reside in shared memory. They are usually written to disk only when the total memory consumption of buffered messages approaches the available shared memory limit.
Note:The portion of a queue that stores buffered messages in memory is sometimes referred to as a buffered queue.
Message retention is not supported for buffered messaging.
When using buffered messaging, Oracle recommends that you do one of the following:
This parameter controls the size of shared memory available to Oracle Streams AQ. If unspecified, up to 10% of the shared pool size may be allocated for the Oracle Streams AQ pool from the database cache.
Turn on SGA autotuning
Oracle will automatically allocate the appropriate amount of memory from the SGA for Oracle Streams AQ, based on Oracle Streams AQ usage as well as usage of other components that use the SGA. Examples of such other components are buffer cache and library cache. If
streams_pool_size is specified, it is used as the lower bound.
See Also:"Setting Initialization Parameters Relevant to Streams" in Oracle Streams Concepts and Administration
This section contains the following topics:
Buffered and persistent messages use the same single-consumer or multiconsumer queues and the same administrative and operational interfaces. They are distinguished from each other by a delivery mode parameter, set by the application when enqueuing the message to an Oracle Streams AQ queue.
See Also:"Enqueuing Messages"
Recipient lists are supported for buffered messaging enqueue.
Buffered messaging is supported in all queue tables created with compatibility 8.1 or higher. Transaction grouping queues and array enqueues are not supported for buffered messages in this release. You can still use the array enqueue procedure to enqueue buffered messages, but the array size must be set to one.
All ordering schemes available for persistent messages are also available for buffered messages, but only within each message class. Ordering among persistent and buffered messages enqueued in the same session is not currently supported.
Both enqueue and dequeue buffered messaging operations must be with
IMMEDIATE visibility mode. Thus they cannot be part of another transaction. You cannot specify delay when enqueuing buffered messages.
Rule-based subscriptions are supported with buffered messaging. The procedure for adding subscribers is enhanced to allow an application to express interest in persistent messages only, buffered messages only, or both.
See Also:"Adding a Subscriber"
Array dequeue is not supported for buffered messaging, but you can still use the array dequeue procedure by setting array size to one message..
Dequeuing applications can choose to dequeue persistent messages only, buffered messages only, or both types. Visibility must be set to
IMMEDIATE for dequeuing buffered messages. All of the following dequeue options are supported:
See Also:"Dequeue Options"
Propagation of buffered messages is supported. A single propagation schedule serves both persistent and buffered messages. The
DBA_QUEUE_SCHEDULES view displays statistics and error information.
Oracle Streams AQ deletes buffered messages once they are propagated to the remote sites. If the receiving site fails before these messages are consumed, then these messages will be lost. The source site will not be able to resend them. Duplicate delivery of messages is also possible.
Oracle Streams AQ implements a flow control system that prevents applications from flooding the shared memory with messages. If the number of unread messages enqueued by a message sender exceeds a system-determined limit, then message sender is blocked until one of the subscribers has read some of its messages. A message sender is identified by
sender_id.name in the enqueue options. A sender blocked due to flow control on a queue does not affect other message senders.
Even with flow control, slow consumers of a multiconsumer queue can cause the number of messages stored in memory to grow without limit. If this happens, older messages are spilled to disk and removed from the Oracle Streams AQ pool to free up memory. This ensures that the cost of disk access is paid by the slower consumers, and faster subscribers can proceed unhindered.
An application can enqueue and dequeue buffered messages from any RAC instance as long as it uses password-based authentication to connect to the database. The structures required for buffered messaging are implemented on one RAC instance. The instance where the buffered messaging structures are implemented is the
OWNER_INSTANCE of the queue table containing the queue. Enqueue and dequeue requests received at other instances are forwarded to the
OWNER_INSTANCE over the interconnect. The
REMOTE_LISTENER parameter in
listener.ora must also be set to enable forwarding of buffered messaging requests to to correct instance.
A service name is associated with each queue in RAC and displayed in the
USER_QUEUES views. This service name always points to the instance with the most efficient access for buffered messaging, minimizing pinging between instances. OCI clients can use the service name for buffered messaging operations.
Oracle recommends that you use buffered messaging with queue-to-queue propagation, a new feature in Oracle Streams AQ 10g Release 2 (10.2). This results in transparent failover when propagating messages to a destination RAC system. You no longer need to repoint your database links if the primary Oracle Streams AQ RAC instance fails.
The following Oracle Streams AQ features are not currently supported for buffered messaging:
Message export and import
Posting for subscriber notification
Retry count and retry delay are not supported for buffered messages. Message expiration is supported. When a buffered message has been in the queue beyond its expiration period, it is moved into the exception queue as a persistent message.
Asynchronous notification allows clients to receive notifications of messages of interest. The client can use these notifications to monitor multiple subscriptions. The client need not be connected to the database to receive notifications regarding its subscriptions. Asynchronous notification is supported for buffered messages. The delivery mode of the message is available in the message descriptor of the notification descriptor.
Note:In releases before Oracle Database 10g Release 2 (10.2), the Oracle Streams AQ notification feature was not supported for queues with names longer than 30 characters. This restriction no longer applies. The 24-character limit on names of user-generated queues still applies. See "Creating a Queue".
The client specifies a callback function which is run for each message. Asynchronous notification cannot be used to invoke an executable, but it is possible for the callback function to invoke a stored procedure.
Clients can receive notifications procedurally using PL/SQL, Java Message Service (JMS), or OCI callback functions, or clients can receive notifications through e-mail or HTTP post. Clients can also specify the presentation for notifications as either
For JMS queues, the dequeue is accomplished as part of the notification; explicit dequeue is not required. For RAW queues, clients can specify payload delivery; but they still must dequeue the message in
REMOVE_NO_DATA mode. For all other persistent queues, the notification contains only the message properties; clients explicitly dequeue to receive the message.
The following notification enhancements are new features of Oracle Streams AQ 10g Release 2 (10.2):
For RAW queues, Oracle Streams AQ clients can now specify that the message payload be delivered along with its notification.
See Also:"AQ Registration Information Type"
In earlier releases of Oracle Streams AQ, message notifications were stored in shared memory and were lost if the instance failed. Clients can now specify persistent message notification. If a RAC instance fails, its notifications are delivered by another RAC node. If a standalone instance fails, its notifications are delivered when the instance restarts.
Note:Notification reliability refers only to server failures. If Oracle Streams AQ is unable to deliver client notifications for any other reason, then the notifications are purged along with the client registration.
See Also:"Notification Quality of Service"
Oracle Streams AQ clients can now use the OCI subscription handle attribute
OCI_ATTR_SUBSCR_PORTNO to designate the port at which notifications are delivered. This is especially useful for clients on a computer behind a firewall. The port for the listener thread can be designated before the first registration, using an attribute in the environment handle. The thread is started the first time an
OCISubscriptionRegister is called. If the client attempts to start another thread on a different port using a different environment handle, then Oracle Streams AQ returns an error.
Note:Designated port notification applies only to OCI clients.
See Also:"Publish-Subscribe Registration Functions in OCI" in Oracle Call Interface Programmer's Guide
In earlier releases of Oracle Streams AQ, registrations for notification persisted until explicitly removed by the client or purged in case of extended client failure. In Oracle Streams AQ 10g Release 2 (10.2) clients can register for a specified time, after which the registration is automatically purged.
When the registration is purged, Oracle Streams AQ sends a notification to the client, so the client can invoke its callback and take any necessary action.
See Also:"AQ Registration Information Type" for information on the
Clients can also register to receive only the first notification, after which the registration is automatically purged.
An example where purge on notification is useful is a client waiting for enqueues to start. In this case, only the first notification is useful; subsequent notifications provide no additional information. Previously, this client would be required to unregister once enqueuing started; now the registration can be configured to go away automatically.
Clients can register for notification of buffered messages. The registration requests apply to both buffered and persistent messages. The message properties delivered with the PL/SQL or OCI notification specify whether the message is buffered or persistent.
Reliable notification is not supported.
The following features apply to enqueuing messages:
When enqueuing messages into a queue, you can operate on an array of messages simultaneously, instead of one message at a time. This can improve the performance of enqueue operations. When enqueuing an array of messages into a queue, each message shares the same enqueue options, but each message can have different message properties. You can perform array enqueue operations using PL/SQL or OCI.
Array enqueuing is not supported for buffered messages in this release.
You can assign an identifier to each message, thus providing a means to retrieve specific messages at a later time.
You can specify the priority of an enqueued message and its exact position in the queue. This means that users can specify the order in which messages are consumed in three ways:
A priority can be assigned to each message.
A sort order specifies which properties are used to order all messages in a queue. This is set when the queue table is created and cannot be changed. You can choose to sort messages by priority, enqueue time, or commit time. The commit-time option, a new feature in Oracle Streams AQ 10g Release 2 (10.2), orders messages by an approximate CSCN calculated for each transaction.
Commit-time ordering is useful when transactions are interdependent or when browsing the messages in a queue must yield consistent results.
A sequence deviation positions a message in relation to other messages.
Note:The sequence deviation feature is deprecated in 10g Release 2 (10.2).
If several consumers act on the same queue, then each consumer gets the first message that is available for immediate consumption. A message that is in the process of being consumed by another consumer is skipped.
Priority ordering of messages is achieved by specifying priority, enqueue time as the sort order. If priority ordering is chosen, then each message is assigned a priority at enqueue time by the enqueuer. At dequeue time, the messages are dequeued in the order of the priorities assigned. If two messages have the same priority, then the order in which they are dequeued is determined by the enqueue time. A first-in, first-out (FIFO) priority queue can also be created by specifying enqueue time, priority as the sort order of the messages.
Messages belonging to one queue can be grouped to form a set that can only be consumed by one user at a time. This requires that the queue be created in a queue table that is enabled for message grouping. All messages belonging to a group must be created in the same transaction, and all messages created in one transaction belong to the same group.
This feature allows users to segment complex messages into simple messages. For example, messages directed to a queue containing invoices can be constructed as a group of messages starting with a header message, followed by messages representing details, followed by a trailer message.
Message grouping is also useful if the message payload contains complex large objects such as images and video that can be segmented into smaller objects.
Group message properties priority, delay, and expiration are determined solely by the message properties specified for the first message in a group, irrespective of which properties are specified for subsequent messages in the group.
The message grouping property is preserved across propagation. However, the destination queue where messages are propagated must also be enabled for transactional grouping. There are also some restrictions you must keep in mind if the message grouping property is to be preserved while dequeuing messages from a queue enabled for transactional grouping.
Applications can mark the messages they send with a custom identification. Oracle Streams AQ also automatically identifies the queue from which a message was dequeued. This allows applications to track the pathway of a propagated message or a string message within the same database.
Messages can be enqueued with an expiration that specifies the interval of time the message is available for dequeuing. The default for expiration is never. When a message expires, it is moved to an exception queue. Expiration processing requires that the queue monitor be running.
The following features apply to dequeuing messages:
When there are multiple processes dequeuing from a single-consumer queue or dequeuing for a single consumer on the multiconsumer queue, different processes skip the messages that are being worked on by a concurrent process. This allows multiple processes to work concurrently on different messages for the same consumer.
A message can be dequeued using one of the following dequeue methods:
Specifying a correlation identifier
A correlation identifier is a user-defined message property. Multiple messages with the same correlation identifier can be present in a queue, which means that the ordering (enqueue order) between messages might not be preserved on dequeue calls.
Specifying a message identifier
Specifying a dequeue condition
A dequeue condition is expressed in terms of message properties or message content and is similar in syntax to the
WHERE clause of a SQL query. Messages in the queue are evaluated against the condition, and messages that satisfy the given condition are returned. When a dequeue condition is used, the order of the messages dequeued is indeterminate, and the sort order of the queue is not honored.
A default dequeue retrieves the first available message.
Note:Dequeuing with correlation identifier, message identifier, or dequeue condition does not preserve the message grouping property.
A dequeue request can browse a message, remove it, or remove it with no data. If a message is browsed, then it remains available for further processing. If a message is removed or removed with no data, then it is no longer available for dequeue requests. Depending on the queue properties, a removed message can be retained in the queue table. A message is retained in the queue table after it has been consumed only if a retention time is specified for its queue.
The browse mode has three risks. First, there is no guarantee that the message can be dequeued again after it is browsed, because a dequeue call from a concurrent user might have removed the message. To prevent a viewed message from being dequeued by a concurrent user, you should view the message in the locked mode.
Second, your dequeue position in browse mode is automatically changed to the beginning of the queue if a nonzero wait time is specified and the navigating position reaches the end of the queue. If you repeat a dequeue call in the browse mode with the
NEXT_MESSAGE navigation option and a nonzero wait time, then you can end up dequeuing the same message over and over again. Oracle recommends that you use a nonzero wait time for the first dequeue call on a queue in a session, and then use a zero wait time with the
NEXT_MESSAGE navigation option for subsequent dequeue calls. If a dequeue call gets an "end of queue" error message, then the dequeue position can be explicitly set by the dequeue call to the beginning of the queue using the
FIRST_MESSAGE navigation option, following which the messages in the queue can be browsed again.
Third, if the sort order of the queue is
PRIORITY, or a combination of these two, then results may not be repeatable from one browse to the next. If you must have consistent browse results, then you should use a commit-time queue.
When a message is dequeued using
REMOVE_NODATA mode, the payload of the message is not retrieved. This mode can be useful when the user has already examined the message payload, possibly by means of a previous
When dequeuing messages from a queue, you can operate on an array of messages simultaneously, instead of one message at a time. This can improve the performance of dequeue operations. If you are dequeuing from a transactional queue, you can dequeue all the messages for a transaction with a single call, which makes application programming easier.
When dequeuing an array of messages from a queue, each message shares the same dequeue options, but each message can have different message properties. You can perform array enqueue and array dequeue operations using PL/SQL or OCI.
Array dequeuing is not supported for buffered messages in this release.
Multiple processes or operating system threads can use the same consumer name to dequeue concurrently from a queue. In that case Oracle Streams AQ provides the first unlocked message that is at the head of the queue and is intended for the consumer. Unless the message identifier of a specific message is specified during dequeue, consumers can dequeue messages that are in the
A message is considered
PROCESSED only when all intended consumers have successfully dequeued the message. A message is considered
EXPIRED if one or more consumers did not dequeue the message before the
EXPIRATION time. When a message has expired, it is moved to an exception queue.
Expired messages from multiconsumer queues cannot be dequeued by the intended recipients of the message. However, they can be dequeued in the
REMOVE mode exactly once by specifying a
NULL consumer name in the dequeue options.
Note:If the multiconsumer exception queue was created in a queue table with the
Queues created in a queue table with
Beginning with Oracle Streams AQ release 8.1.6, only the queue monitor removes messages from multiconsumer queues. This allows dequeuers to complete the dequeue operation by not locking the message in the queue table. Because the queue monitor removes messages that have been processed by all consumers from multiconsumer queues approximately once every minute, users can see a delay between when the messages have been completely processed and when they are physically removed from the queue.
You have several options for selecting a message from a queue. You can select the first message with the
FIRST_MESSAGE navigation option. Alternatively, once you have selected a message and established its position in the queue, you can then retrieve the next message with the
NEXT_MESSAGE navigation option.
FIRST_MESSAGE navigation option performs a
SELECT on the queue. The
NEXT_MESSAGE navigation option fetches from the results of the
SELECT run in the
FIRST_MESSAGE navigation. Thus performance is optimized because subsequent dequeues need not run the entire
If the queue is enabled for transactional grouping, then the navigation options work in a slightly different way. If
FIRST_MESSAGE is requested, then the dequeue position is still reset to the beginning of the queue. But if
NEXT_MESSAGE is requested, then the position is set to the next message in the same transaction. Transactional grouping also offers a
NEXT_TRANSACTION option. It sets the dequeue position to the first message of the next transaction.
If you reach the end of the queue while using the
NEXT_TRANSACTION option, and you have specified a nonzero wait time, then the navigating position is automatically changed to the beginning of the queue. If a zero wait time is specified, then you can get an exception when the end of the queue is reached.
Oracle Streams AQ allows applications to block on one or more queues waiting for the arrival of either a newly enqueued message or a message that becomes ready. You can use the
DEQUEUE operation to wait for the arrival of a message in a single queue or the
LISTEN operation to wait for the arrival of a message in more than one queue.
Note:Applications can also perform a blocking dequeue on exception queues to wait for arrival of
When the blocking
DEQUEUE call returns, it returns the message properties and the message payload. When the blocking
LISTEN call returns, it discloses only the name of the queue where a message has arrived. A subsequent
DEQUEUE operation is needed to dequeue the message.
When there are messages for multiple agents in the agent list,
LISTEN returns with the first agent for whom there is a message. To prevent one agent from starving other agents for messages, the application can change the order of the agents in the agent list.
Note:This feature is not currently supported in Visual Basic (OO4O).
Applications can optionally specify a timeout of zero or more seconds to indicate the time that Oracle Streams AQ must wait for the arrival of a message. The default is to wait forever until a message arrives in the queue. This removes the burden of continually polling for messages from the application, and it saves CPU and network resources because the application remains blocked until a new message is enqueued or becomes
READY after its
An application that is blocked on a dequeue is either awakened directly by the enqueuer if the new message has no
DELAY or is awakened by the queue monitor process when the
EXPIRATION time has passed. If an application is waiting for the arrival of a message in a remote queue, then the Oracle Streams AQ propagator wakes up the blocked dequeuer after a message has been propagated.
If the transaction dequeuing a message from a queue fails, then it is regarded as an unsuccessful attempt to consume the message. Oracle Streams AQ records the number of failed attempts to consume the message in the message history. Applications can query the
RETRY_COUNT column of the queue table view to find out the number of unsuccessful attempts on a message. In addition, Oracle Streams AQ allows the application to specify, at the queue level, the maximum number of retries for messages in the queue. The default value for maximum retries is 5. If the number of failed attempts to remove a message exceeds this number, then the message is moved to the exception queue and is no longer available to applications.
Note:If a dequeue transaction fails because the server process dies (including
A bad condition can cause the transaction receiving a message to end. Oracle Streams AQ allows users to hide the bad message for a specified retry delay interval, during which it is in the WAITING state. After the retry delay, the failed message is again available for dequeue. The Oracle Streams AQ time manager enforces the retry delay property. The default value for retry delay is 0.
If multiple sessions are dequeueing messages from a queue simultaneously, then
RETRY_COUNT information might not always be updated correctly. If session one dequeues a message and rolls back the transaction, then Oracle Streams AQ notes that the
RETRY_COUNT information for this message must be updated. However
RETRY_COUNT cannot be incremented until session one completes the rollback. If session two attempts to dequeue the same message after session one has completed the rollback but before it has incremented
RETRY_COUNT, then the dequeue by session two succeeds. When session one attempts to increment
RETRY_COUNT, it finds that the message is locked by session two and
RETRY_COUNT is not incremented. A trace file is then generated in the user_dump_destination for the instance with the following message:
Error on rollback: ORA-25263: no message in queue schema.qname with message ID ...
Note:Maximum retries and retry delay are not available with 8.0-style multiconsumer queues.
Queues created in a queue table with
Enqueue and dequeue requests are usually part of a transaction that contains the requests, thereby providing the wanted transactional action. You can, however, specify that a specific request is a transaction by itself, making the result of that request immediately visible to other transactions. This means that messages can be made visible to the external world when the enqueue or dequeue statement is applied or after the transaction is committed.
Note:Transaction protection is not supported for buffered messaging.
An exception queue is a repository for expired or unserviceable messages. Applications cannot directly enqueue into exception queues. Also, a multiconsumer exception queue cannot have subscribers associated with it. However, an application that intends to handle these expired or unserviceable messages can dequeue them exactly once from the exception queue using remove mode. The consumer name specified while dequeuing should be null. Messages can also be dequeued from the exception queue by specifying the message identifier.
Note:Expired or unserviceable buffered messages are moved to an exception queue as persistent messages.
Messages intended for single-consumer queues, or for 8.0-style multiconsumer queues, can only be dequeued by their message identifiers once the messages have been moved to an exception queue.
Queues created in a queue table with
After a message has been moved to an exception queue, there is no way to identify which queue the message resided in before moving to the exception queue. If this information is important, then the application must save this information in the message itself.
The exception queue is a message property that can be specified during enqueue time. If an exception queue is not specified, then a default exception queue is used. The default exception queue is automatically created when the queue table is created.
A message is moved to an exception queue under the following conditions:
It was not dequeued within the specified expiration interval.
For a message intended for multiple recipients, the message is moved to the exception queue if one or more of the intended recipients was not able to dequeue the message within the specified expiration interval. The default expiration interval is never, meaning the messages does not expire.
The message was dequeued successfully, but the application that dequeued it rolled back the transaction because of an error that arose while processing the message. If the message has been dequeued but rolled back more than the number of times specified by the retry limit, then the message is moved to the exception queue.
For a message intended for multiple recipients, a separate retry count is kept for each recipient. The message is moved to the exception queue only when retry counts for all recipients of the message have exceeded the specified retry limit.
The default retry limit is five for single-consumer queues and 8.1-style multiconsumer queues. No retry limit is supported for 8.0-style multiconsumer queues, which are deprecated in Oracle Streams AQ 10g Release 2 (10.2)..
Note:If a dequeue transaction fails because the server process dies (including
The statement processed by the client contains a dequeue that succeeded but the statement itself was undone later due to an exception.
If the dequeue procedure succeeds but the PL/SQL procedure raises an exception, then Oracle Streams AQ increments the retry count of the message returned by the dequeue procedure.
The client program successfully dequeued a message but terminated before committing the transaction.
Messages can be propagated from one queue to another, allowing applications to communicate with each other without being connected to the same database or to the same queue. The destination queue can be located in the same database or in a remote database.
Propagation enables you to fan out messages to a large number of recipients without requiring them all to dequeue messages from a single queue. You can also use propagation to combine messages from different queues into a single queue. This is known as compositing or funneling messages.
Note:You can propagate messages from a multiconsumer queue to a single-consumer queue. Propagation from a single-consumer queue to a multiconsumer queue is not possible.
A message is marked as processed in the source queue immediately after the message has been propagated, even if the consumer has not dequeued the message at the remote queue. Similarly, when a propagated message expires at the remote queue, the message is moved to the exception queue of the remote queue, and not to the exception queue of the local queue. Oracle Streams AQ does not currently propagate the exceptions to the source queue.
To enable propagation, one or more subscribers are defined for the queue from which messages are to be propagated and a schedule is defined for each destination where messages are to be propagated from the queue. Oracle Streams AQ automatically checks if the type of the remote queue is structurally equivalent to the type of the local queue. Messages enqueued in the source queue are then propagated and automatically available for dequeuing at the destination queue or queues.
When messages arrive at the destination queues, sessions based on the source queue schema name are used for enqueuing the newly arrived messages into the destination queues. This means that you must grant schemas of the source queues enqueue privileges to the destination queues.
Propagation is performed by two or more job queue background processes. This is in addition to the number of job queue background processes needed for handling jobs unrelated to propagation.
Oracle Streams AQ offers two kinds of propagation:
A single propagation schedule is used to propagate messages to all subscribing queues. Hence any changes made to this schedule will affect message delivery to all the subscribing queues.
Queue-to-queue propagation delivers messages or events from the source queue to a specific destination queue identified on the dblink. This allows the user to have fine-grained control on the propagation schedule for message delivery.
This new propagation mode also supports transparent failover when propagating to a destination RAC system. With queue-to-queue propagation, you are no longer required to repoint a database link if the owner instance of the queue fails on RAC.
Oracle Streams AQ provides detailed statistics about the messages propagated and the schedule itself. This information can be used to tune propagation schedules for best performance.
Consumers of a message in multiconsumer queues can be local or remote. Local consumers dequeue messages from the same queues into which the producer enqueued the messages. Local consumers have a name but no address or protocol in their agent descriptions.
Remote consumers dequeue from queues that are different from the queues where the messages were enqueued. Remote consumers fall into three categories:
The address refers to a queue in the same database.
In this case the consumer dequeues the message from a different queue in the same database. These addresses are of the form
queue_name. If the schema is not specified, then the schema of the current user is used.
The address refers to a queue in a different database.
In this case the database must be reachable using database links and the protocol must be either
NULL or 0. These addresses are of the form
queue_name@dblink. If the schema is not specified, then the schema of the current user is used. If the database link does not have a domain name specified, then the default domain as specified by the
ora parameter is used.
The address refers to a destination that can be reached by a third party protocol.
You must refer to the documentation of the third party software to determine how to specify the address and the protocol database link and schedule propagation.
The delay, expiration, and priority parameters apply identically to both local and remote consumers in both queue-to-dblink and queue-to-queue propagation. Oracle Streams AQ accounts for any delay in propagation by adjusting the delay and expiration parameters accordingly. For example, if expiration is set to one hour, and the message is propagated after 15 minutes, then the expiration at the remote queue is set to 45 minutes.
Figure 1-9 illustrates applications on different databases communicating using Oracle Streams AQ. Each application has an inbox for handling incoming messages and an outbox for handling outgoing messages. Whenever an application enqueues a message, it goes into its outbox regardless of the message destination. Similarly, an application dequeues messages from its inbox no matter where the message originates.
A queue-to-dblink propagation schedule is defined for a pair of source and destination database links. A queue-to-queue propagation schedule is defined for a pair of source and destination queues. If a queue has messages to be propagated to several queues, then a schedule must be defined for each of the destination queues. With queue-to-dblink propagation, all schedules for a particular remote database have the same frequency. With queue-to-queue propagation, the frequency of each schedule can be adjusted independently of the others
A schedule indicates the time frame during which messages can be propagated from the source queue. This time frame can depend on a number of factors such as network traffic, load at the source database, and load at the destination database. If the duration is unspecified, then the time frame is an infinite single window. If a window must be repeated periodically, then a finite duration is specified along with a
next_time function that defines the periodic interval between successive windows.
When a schedule is created, a job is automatically submitted to the job queue facility to handle propagation.
The propagation schedules defined for a queue can be changed or dropped at any time during the life of the queue. You can also temporarily disable a schedule instead of dropping it. All administrative calls can be made irrespective of whether the schedule is active or not. If a schedule is active, then it takes a few seconds for the calls to be processed.
Large Objects can be propagated using Oracle Streams AQ using two methods:
RAW queues the message payload is stored as a BLOB. This allows users to store up to 32KB of data when using the PL/SQL interface and as much data as can be contiguously allocated by the client when using OCI. This method is supported by all releases after 8.0.4 inclusive.
Propagation from object queues with LOB attributes
The user can populate the
LOB and read from the
LOB using Oracle Database
LOB handling routines. The
LOB attributes can be
CLOBs (not NCLOBs). If the attribute is a
CLOB, then Oracle Streams AQ automatically performs any necessary characterset conversion between the source queue and the destination queue. This method is supported by all releases from 8.1.3 inclusive.
Note:Payloads containing LOBs require users to grant explicit
Detailed runtime information about propagation is gathered and stored in the
DBA_QUEUE_SCHEDULES view for each propagation schedule. This information can be used by queue designers and administrators to fix problems or tune performance. Similarly, errors reported by the view can be used to diagnose and fix problems. The view also describes additional information such as the session ID of the session handling the propagation and the process name of the job queue process handling the propagation.
For each schedule, detailed propagation statistics are maintained:
Total number of messages propagated in a schedule
Total number of bytes propagated in a schedule
Maximum number of messages propagated in a window
Maximum number of bytes propagated in a window
Average number of messages propagated in a window
Average size of propagated messages
Average time to propagated a message
Propagation has built-in support for handling failures and reporting errors. For example, if the specified database link is invalid, if the remote database is unavailable, or if the remote queue is not enabled for enqueuing, then the appropriate error message is reported. Propagation uses an exponential backoff scheme for retrying propagation from a schedule that encountered a failure.
If a schedule continuously encounters failures, then the first retry happens after 30 seconds, the second after 60 seconds, the third after 120 seconds and so forth. If the retry time is beyond the expiration time of the current window, then the next retry is attempted at the start time of the next window. A maximum of 16 retry attempts is made, after which the schedule is automatically disabled.
Note:Once a retry attempt slips to the next propagation window, it will always do so; the exponential backoff scheme no longer governs retry scheduling. If the date function specified in the
When a schedule is disabled automatically due to failures, the relevant information is written into the alert log. A check for scheduling failures indicates:
How many successive failures were encountered
The error message indicating the cause for the failure
The time at which the last failure was encountered
By examining this information, a queue administrator can fix the failure and enable the schedule. If propagation is successful during a retry, then the number of failures is reset to 0.
In some situations that indicate application errors in queue-to-dblink propagations, Oracle Streams AQ marks messages as
UNDELIVERABLE and logs a message in
alert.log. Examples of such errors are when the remote queue does not exist or when there is a type mismatch between the source queue and the remote queue. The trace files in the
background_dump_dest directory can provide additional information about the error.
When a new job queue process starts, it clears the mismatched type errors so the types can be reverified. If you have capped the number of job queue processes and propagation remains busy, then you might not want to wait for the job queue process to terminate and restart. Queue types can be reverified at any time using
Note:When a type mismatch is detected in queue-to-queue propagation, propagation stops and throws an error. In such situations you must query the
Propagation has support built-in for Oracle Real Application Clusters. It is transparent to the user and the queue administrator. The job that handles propagation is submitted to the same instance as the owner of the queue table where the queue resides.
If there is a failure at an instance and the queue table that stores the source queue is migrated to a different instance, then the propagation job is also migrated to the new instance. This minimizes pinging between instances and thus offers better performance.
The destination can be identified by a database link or by destination queue name. Specifying the destination database results in queue-to-dblink propagation. If you propagate messages to several queues in another database, then all queue-to-dblink propagations to that database have the same frequency. Specifying the destination queue name results in queue-to-queue propagation, a new feature in Oracle Streams AQ 10g Release 2 (10.2). If you propagate messages to several queues in another database, then queue-to-queue propagation enables you to adjust the frequency of each schedule independently of the others. You can even enable or disable individual propagations.
This new queue-to-queue propagation mode also supports transparent failover when propagating to a destination RAC system. With queue-to-queue propagation, you are no longer required to repoint a database link if the owner instance of the queue fails on RAC.
See Also:"Scheduling a Queue Propagation" for more information on queue-to-queue propagation
Propagation has been designed to handle any number of concurrent schedules. The number of job queue processes is limited to a maximum of 1000, and some of these can be used to handle jobs unrelated to propagation. Hence, propagation has built-in support for multitasking and load balancing.
The propagation algorithms are designed such that multiple schedules can be handled by a single job queue process. The propagation load on a job queue process can be skewed based on the arrival rate of messages in the different source queues.
If one process is overburdened with several active schedules while another is less loaded with many passive schedules, then propagation automatically redistributes the schedules so they are loaded uniformly.
If the protocol number for a recipient is in the range 128 - 255, then the address of the recipient is not interpreted by Oracle Streams AQ and the message is not propagated by the Oracle Streams AQ system. Instead, a third-party propagator can dequeue the message by specifying a reserved consumer name in the dequeue operation. The reserved consumer names are of the form
protocol_number. For example, the consumer name
AQ$_P128 can be used to dequeue messages for recipients with protocol number 128. The list of recipients for a message with the specific protocol number is returned in the
recipient_list message property on dequeue.
Another way for Oracle Streams AQ to propagate messages to and from third-party messaging systems is through Messaging Gateway. Messaging Gateway dequeues messages from an Oracle Streams AQ queue and guarantees delivery to supported third-party messaging systems. Messaging Gateway can also dequeue messages from these systems and enqueue them to an Oracle Streams AQ queue.
In Oracle Database 10g you can set up Oracle Streams AQ propagation over HTTP and HTTPS (HTTP over SSL). HTTP propagation uses the Internet access infrastructure and requires that the Oracle Streams AQ servlet that connects to the destination database be deployed. The database link must be created with the connect string indicating the Web server address and port and indicating HTTP as the protocol. The source database must be created for running Java and XML. Otherwise, the setup for HTTP propagation is more or less the same as Oracle Net Services propagation.
Applications often use data in different formats. A transformation defines a mapping from one Oracle data type to another. The transformation is represented by a SQL function that takes the source data type as input and returns an object of the target data type. Only one-to-one message transformations are supported.
To transform a message during enqueue, specify a mapping in the enqueue options. To transform a message during dequeue, specify a mapping either in the dequeue options or when you add a subscriber. A dequeue mapping overrides a subscriber mapping. To transform a message during propagation, specify a mapping when you add a subscriber.
You can create transformations by creating a single PL/SQL function or by creating an expression for each target type attribute. The PL/SQL function returns an object of the target type or the constructor of the target type. This representation is preferable for simple transformations or those not easily broken down into independent transformations for each attribute.
Creating a separate expression specified for each attribute of the target type simplifies transformation mapping creation and management for individual attributes of the destination type. It is useful when the destination type has many attributes.
As Figure 1-10 shows, queuing, routing, and transformation are essential building blocks to an integrated application architecture. The figure shows how data from the Out queue of a CRM application is routed and transformed in the integration hub and then propagated to the In queue of the Web application. The transformation engine maps the message from the format of the Out queue to the format of the In queue.
You can transform XML data using the
extract() method supported on
XMLType to return an object of
XMLType after applying the supplied
XPath expression. You can also create a PL/SQL function that transforms the
XMLType object by applying an XSLT transformation to it, using the package
This section contains these topics:
Before 10g Release 1 (10.1), the Oracle Streams AQ time manager process was called queue monitor (
QMNn), a background process controlled by setting the dynamic
AQ_TM_PROCESSES. Beginning with 10g Release 1 (10.1), time management and many other background processes are automatically controlled by a coordinator-slave architecture called Queue Monitor Coordinator (
QMNC dynamically spawns slaves named
qXXX depending on the system load. The slaves provide mechanisms for:
Garbage collection for the queue table
Memory management tasks for buffered messages
Because the number of processes is determined automatically and tuned constantly, you are saved the trouble of setting it with
Although it is no longer necessary to set
AQ_TM_PROCESSES, it is still supported. If you do set it (up to a maximum of 10), then QMNC still autotunes the number of processes. But you are guaranteed at least the set number of processes for persistent queues. Processes for a buffered queue and other Oracle Streams tasks, however, are not affected by this parameter.
Note:If you want to disable the Queue Monitor Coordinator, then you must set
Oracle Internet Directory is a native LDAPv3 directory service built on Oracle Database that centralizes a wide variety of information, including e-mail addresses, telephone numbers, passwords, security certificates, and configuration data for many types of networked devices. You can look up enterprise-wide queuing information—queues, subscriptions, and events—from one location, the Oracle Internet Directory. Refer to the Oracle Internet Directory Administrator's Guide for more information.
You can use Oracle Enterprise Manager to:
Create and manage queues, queue tables, propagation schedules, and transformations
Monitor your Oracle Streams AQ environment using its topology at the database and queue levels, and by viewing queue errors and queue and session statistics
The systems administrator specifies the retention duration to retain messages after consumption. Oracle Streams AQ stores information about the history of each message, preserving the queue and message properties of delay, expiration, and retention for messages destined for local or remote receivers. The information contains the enqueue and dequeue times and the identification of the transaction that executed each request. This allows users to keep a history of relevant messages. The history can be used for tracking, data warehouse, and data mining operations, as well as specific auditing functions.
Message retention is not supported for buffered messaging.
The Oracle Streams AQ retention feature can be used to automatically clean up messages after the user-specified duration after consumption.
If messages are accidentally inserted into a queue for the wrong subscriber, you can dequeue them with the subscriber name or by message identifier. This consumes the messages, which are cleaned up after their retention time expires.
To clean up messages for a particular subscriber, you can remove the subscriber and add the subscriber again. Removing the subscriber removes all the messages for that subscriber.
Retained messages can be related to each other to form sequences. These sequences represent event journals, which are often constructed by applications. Oracle Streams AQ is designed to let applications create event journals automatically.
Oracle Streams AQ maintains the entire history of information about a message along with the message itself. This information serves as proof of sending and receiving of messages and can be used for nonrepudiation of the sender and nonrepudiation of the receiver.
The following information is kept at enqueue for nonrepudiation of the enqueuer:
Oracle Streams AQ agent doing the enqueue
Database user doing the enqueue
Transaction ID of the transaction doing enqueue
The following information is kept at dequeue for nonrepudiation of the dequeuer:
Oracle Streams AQ agent doing dequeue
Database user doing dequeue
Transaction ID of the transaction doing dequeue
After propagation, the
Original_Msgid field in the destination queue of the propagation corresponds to the message ID of the source message. This field can be used to correlate the propagated messages. This is useful for nonrepudiation of the dequeuer of propagated messages.
Stronger nonrepudiation can be achieved by enqueuing the digital signature of the sender at the time of enqueue with the message and by storing the digital signature of the dequeuer at the time of dequeue.
You can access Oracle Streams AQ over the Internet by using Simple Object Access Protocol (SOAP). Internet Data Access Presentation (IDAP) is the SOAP specification for Oracle Streams AQ operations. IDAP defines the XML message structure for the body of the SOAP request.
An IDAP message encapsulates the Oracle Streams AQ request and response in XML. IDAP is used to perform Oracle Streams AQ operations such as enqueue, dequeue, send notifications, register for notifications, and propagation over the Internet standard transports—HTTP(s) and e-mail. In addition, IDAP encapsulates transactions, security, transformation, and the character set ID for requests.
You can create an alias to an Oracle Streams AQ agent in Oracle Internet Directory and then use the alias in IDAP documents sent over the Internet to perform Oracle Streams AQ operations. Using aliases prevents exposing the internal name of the Oracle Streams AQ agent.
Figure 1-11 shows the architecture for performing Oracle Streams AQ operations over HTTP. The major components are:
Oracle Streams AQ client program
Web server/servlet runner hosting the Oracle Streams AQ servlet
Oracle Database server
The Oracle Streams AQ client program sends XML messages (conforming to IDAP) to the Oracle Streams AQ servlet, which understands the XML message and performs Oracle Streams AQ operations. Any HTTP client, a Web browser for example, can be used. The Web server/servlet runner hosting the Oracle Streams AQ servlet, Apache/Jserv or Tomcat for example, interprets the incoming XML messages. The Oracle Streams AQ servlet connects to the Oracle Database server and performs operations on user queues.
Note:This feature is certified to work with Apache, along with the Tomcat or Jserv servlet execution engines. However, the code does not prevent the servlet from working with other Web server and servlet execution engines that support Java Servlet 2.0 or higher interfaces.
You can access Oracle Streams AQ functionality through the following interfaces:
Visual Basic using Oracle Objects for OLE
Java Message Service (JMS) using the
oracle.jms Java package
Internet access using HTTP(S)
Oracle Streams AQ demos can be installed from the Oracle Database Companion CD. Once they are installed, you can find them in the
$ORACLE_HOME/rdbms/demo directory. Refer to
aqjmsREADME.txt in the
demo directory for more information.
|Demo and Locations||Topic|
||Create users, message types, and tables|
||Create queue tables, queues, subscribers, and propagation schedule|
||Enqueue messages into input queue|
||Install dequeue procedures|
||Perform blocking dequeues|
||Perform listen for multiple agents|
||Clean up users, queue tables, queues, and subscribers in
||Enqueue and dequeue to XMLType queue using XPATH expressions|
||Demonstrates server-to-server email notifications with default XML presentation|
||Set up queues and subscribers for array enqueue and dequeue (for OCI array demos also)|
||Array enqueue 10 messages|
||Array dequeue 10 messages|
||Clean up queues and subscribers for array enqueue and dequeue (for OCI array demos also)|
||Perform blocking dequeues|
||Perform listen for multiple agents|
||Array enqueue 10 messages|
||Array dequeue 10 messages|
|Demo and Locations||Topic|
||Describes the Oracle Streams AQ Java API and JMS demos|
||Set up Oracle Streams AQ JMS demos|
||Enqueue text messages and dequeue based on message properties|
||Message listener demo (enqueue messages)|
||Message listener demo (set up listener and dequeue messages)|
||Oracle type payload: dequeue on payload content|
||Queue browser example|
||Schedule propagation between queues in the database|
||Send and receive an ADT message containing XML data|
||JMS 1.1 domain unification demo|
||JMS bulk array enqueue and dequeue|
||ANYDATA messaging with JMS message types and ADT messages|
||Clean up AQ JMS demos|
||Set up Oracle Streams AQ Java API demos|
||Enqueue and dequeue RAW messages|
||Enqueue and dequeue object type messages using ORAData interface|
||Clean up AQ Java API demos|
||Enqueues and dequeues a message within the database|
||Set up kprb driver demo|
||Defines Java program
||Clean up AQ kprb driver demo|
|Demo and Locations||Topic|
||Describes the Internet access demos|
||Create users, queue tables, and queues|
||AQXmlSend: Enqueue three messages to an ADT single- consumer queue with piggyback commit|
||AQXmlReceive: Dequeue messages from ADT single-consumer queue with piggyback commit|
||AQXmlPublish: Enqueue two messages to an ADT multiconsumer queue|
||AQXmlReceive: Dequeue messages from an ADT (with LOB) multiconsumer queue|
||AQXmlCommit: Commit previous operation|
||AQXmlSend: Enqueue a message to a JMS TEXT single-consumer queue with piggyback commit|
||AQXmlReceive: Dequeue messages from a JMS TEXT single-consumer queue with piggyback commit|
||AQXmlPublish: Enqueue a JMS MAP message with recipient into multiconsumer queue|
||AQXmlReceive: Dequeue JMS MAP messages from a multiconsumer queue|
||AQXmlRollback: Roll back previous operation|
||Servlet to post Oracle Streams AQ XML files (for Jserv)|
||Servlet for Oracle Streams AQ HTTP propagation|
||Clean up AQ XML demo|