I have difficulties wrapping my head around the concept.
I am trying to implement an endpoint that listens on a tcp port for incoming messages in a proprietary format, which would then transform the message and have camel take over the forwarding and routing.
Did I understand correctly that that the Producer is responsible for sending messages into the Endpoint and the Consumer receives them from that endpoint?
When studying the interfaces I couldn’t figure out the message flow between those objects, especially on the consumer part. Consumer only defines start() and stop() methods…
When setting up a test on a skeleton implementation, Camel invoked createProducer() on the endpoint and process() on the producer object. After that it returned, swithout doing anything with the consumer or the processor associated with it.
Could someone point me in the right direction?
I finally figured it out by looking at the Stream component.
Turns out that I made the mistake of thinking about the endpoint as something central thru which everything must go.
The simple answer is that the consumer receives data from an external system (listening on a server socket in my case) and the producer sends data to the external system.
Since my endpoint is read-only (it will not be used as the final destination of Camel routing process), I really do not need a producer (it should throw a RuntimeException if the system still tries to do it due to misconfiguration). A fitting example would be the camel-atom endpoint – you can read feeds but (as of 1.6.0) you cannot publish one.
Likewise, you only need a producer for a write-only endpoint which does not receive data from an external system (e.g logging).