Building ActiveMQ networks of brokers is a handy way to eliminate the need of making application deal with network outage. If the application lives on the same server or highly reliable network as the broker, then there is no need to implement network based error handling. However often consumers are remote and they also do not want to deal with network connectivity and those consumers may choose to implement their own broker. Connecting brokers then allows brokers to deal with intermittent networks and transmit messages whenever possible while exposing always on interface to local customers.

In his new tutorial @ripienaar covers how how to build ActiveMQ networks of brokers, so check it out.