Currently, with linkRoute mode, when there are network glitches between AMQ Interconnect Router and AMQ Brokers, the link could be re-activated, but the end-end connections from JMS clients are still affected if the producer has been cached by the Connection Factory. 

To improve the self-healing of the connections with frontend client, propose below solutions:

1. Whenever the processing broker instance (1 out of 3 in the broker cluster) has network issues, interconnect should be able to automatically switch to other healthy brokers (1 from the rest 2) and frontend client should be not aware of it. The old connection established by the client should be maintained at always and no need to restart client.
2. Whenever the Interconnect router detect network glitches to the processing broker instance, it automatically kill the current client connection and it is up to the client side to re-establish a new connection (In Spring, this is controlled by ccf.setReconnectiononException(true), which is true by default)

This is to enhance the Interconnect router connection self-healing with frontend client.