[rabbitmq-discuss] RabbitMQ Deployment Architectures (for lack of a better subject)

[Michael Klishin]

On 12 April 2014 at 21:56:40, Greg Poirier (greg.poirier at opower.com) wrote:
> > So this would require, for example, the consumers to connect  
> directly to the storage clusters

Yes. I thought the point was for consumers to use these “storage” clusters.

 in order to instantiate their  
> queues and then federate those queues to exchanges on the router  
> systems?

Consumers won’t federate anything. They will set up the routing as if the publishers
were publishing to their local “storage” cluster. This will require them knowing
only what exchange gets what messages and nothing else. This is pretty straightforward.

The hard part would be replying to publishers that publish to a different cluster.
This is something that’s currently not addressed by the Federation plugin at all.
But in a fair share of cases communication between publishers and consumers is uni-directional. 
--  
MK  

Software Engineer, Pivotal/RabbitMQ