activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Petrucci Andreas <>
Subject RE: Distributed ActiveMQ Brokers
Date Mon, 27 Dec 2010 14:53:28 GMT

thanks for the reply, i'm aware of the network-of-brokers solution but i doubt if 2 distinct
producers can produce messages for the SAME queue at 2 different brokers of the network. Yes,
in my setup i really need multiple brokers but i think that what i ask for (i think that 
it) cannot be provided by a network of brokers .

> To:;
> From:
> Subject: Re: Distributed ActiveMQ Brokers
> Date: Mon, 27 Dec 2010 15:31:03 +0100
> Afternoon,
> As far as I understand your use case, you can set up a so called network of brokers to
achieve this.
> A NWOB is a collection of brokers that know about each other. Destinations are distributed
on the participating nodes as far as they're involved. Clients can connect to any of the nodes
to produce messages. The same applies for consumption, since messages are forwarded to one
of the nodes that has a subscriber for a particular destination.
> But: You wrote about multiple brokers within one network. You should double check if
you really need multiple brokers, since the main reason for a NWOB would be to scale horizontally
(for queues) or to distribute destinations over multiple networks. Unless you really have
significant traffic, one node can handle quite a lot of queue messages (several k/s, depending
on infrastructure and persistence).
> Hope that helps.
> Dirk Fröhner
> Sent from one of my smartphones
> ----- Reply message -----
> From: "Petrucci Andreas" <>
> Date: Mon, Dec 27, 2010 12:50
> Subject: Distributed ActiveMQ Brokers
> To: <>
> Hello there, all i want ot ask is there there is a specific setup of activemq brokers
such that we will have multiple brokers in our network that can serve clients that are producing
(just producing) messages for one ( just one) queue. As far as i can see , there is no way
to achieve that. Am I right? Any ideas? :)
> thnks in advance,
> Petrucci
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message