Architectures based on a single queue manager
The simplest IBM MQ architectures involve the configuration and use of a single queue manager.
Before we plan the IBM MQ architecture, familiarize yourself with the basic IBM MQ concepts. See IBM MQ Technical overview.
A number of possible architectures using a single queue manager are described in the following sections:
- Single queue manager with local applications accessing a service
- Single queue manager with remote applications accessing a service as clients
- Single queue manager with a publish/subscribe configuration
Single queue manager with local applications accessing a service
The first architecture based on a single queue manager is where the applications accessing a service are running on the same system as the applications providing the service. An IBM MQ queue manager provides asynchronous intercommunication between the applications requesting the service and the applications providing the service. This means that communication between the applications can continue even if one of the applications is offline for an extended period of time.
Single queue manager with remote applications accessing a service as clients
The second architecture based on a single queue manager has the applications running remotely from the applications providing the service. The remote applications are running on different systems to the services. The applications connect as clients to the single queue manager. This means that access to a service can be provided to multiple systems through a single queue manager.
A limitation of this architecture is that a network connection must be available for an application to operate. The interaction between the application and the queue manager over the network connection is synchronous.
Single queue manager with a publish/subscribe configuration
An alternative architecture using a single queue manager is to use a publish/subscribe configuration. In publish/subscribe messaging, we can decouple the provider of information from the consumers of that information. This differs from the point to point messaging styles in the previously described architectures, where the applications must know information about the target application, for example the queue name to put messages on. Using IBM MQ publish/subscribe the sending application publishes a message with a specified topic based on the subject of the information. IBM MQ handles the distribution of the message to applications that have registered an interest in that subject through a subscription. The receiving applications also do not need to know anything about the source of the messages to receive them. For more information, see Publish/subscribe messaging and Example of a single queue manager publish/subscribe configuration.
Parent topic: Plan an IBM MQ architecture
Related tasks
Related information