[rabbitmq-discuss] Broker to Broker communication...

Jeff Bannister jeff at gamesys.co.uk
Mon Aug 9 20:54:47 BST 2010



Matthias Radestock-3 wrote:
> 
> 
> rabbitmq-shovel is certainly one option. But do you actually need two 
> brokers in the first place? Why not co-locate the broker with either the 
> sending or the receiving side?
> 
> 

I have no preference or reason for having a broker each side. This is the
conclusion I came to after reading through some docs. I may well lack some
fundamental understanding of AMQP/RabbitMQ. With the approach you suggest I
am not sure how this would work. 

The application that produces the messages would send them asynchronously to
a local broker i.e. an Exchange. It is my understanding that an Exchange
will not store any messages, therefore I am guessing I need a Q on the
sending side to buffer these messages. This would give me an Exchange and a
Q on the sending side. How would the messages get sent to the destination
without something like shovel? i.e. Who would read the Q, the consumer on
the other side of the WAN?

So I would end up have something like...

App A-->Exchange-->Q-->+WAN+>Consumer-->AppB

I am keen to have the simplest solution I can. Therefore if 'shovel' is an
unnecessary part then I would be happy to loose it.



-- 
View this message in context: http://old.nabble.com/Broker-to-Broker-communication...-tp29390117p29391546.html
Sent from the RabbitMQ mailing list archive at Nabble.com.



More information about the rabbitmq-discuss mailing list