On Fri, Feb 20, 2009 at 9:22 PM, David Ingham <[email protected]> wrote:
> I'm struggling to think of a business case for why I'd want to mix and match > C++ and Java brokers in a cluster. I'm not pushing back; I'm just interested > as to why you think this is important. It's not so much important, more interesting. I think one case when I would have a heterogeneous cluster would be where I've had a system grow organically and there's some reason why the extant brokers can't just be swapped out. The only time I think I would purposefully design a system with a mix (which is clearly somewhat obtuse) is where I'm hedging against some catastrophic bug that affects, say, the java broker but not the c++ broker and the system absolutely, positively cannot cope. > FWIW, I think that there's great value in having a common clustering design > across the different language brokers, mainly to reduce overall complexity of > the project as a whole, but I hadn't considered it being a requirement to mix > different language versions in the same cluster. It's definitely not something I'm saying is a must-have, I just think it's sufficiently nice to have that it's worth trying not to preclude it. - Aidan -- Apache Qpid - World Domination through Advanced Message Queueing http://qpid.apache.org --------------------------------------------------------------------- Apache Qpid - AMQP Messaging Implementation Project: http://qpid.apache.org Use/Interact: mailto:[email protected]
