[zeromq-dev] design for peer to peer messaging

Pieter Hintjens ph at imatix.com
Sat Mar 9 12:49:01 CET 2013


In general it's not useful to design a 0MQ architecture until you've
fully internalized how the different patterns work.

My advice is to make a *minimal* working design with one of each
piece, add more one piece at a time, and work out the issues like
that. When you've done this a few times it'll be obvious.


On Sat, Mar 9, 2013 at 11:18 AM, Michael Fung <mike at 3open.org> wrote:
> Dear all,
> I would like to build a service for electronics hobbyists to control
> their devices from anywhere. The targeted devices are primitive and
> probably cannot speak 0MQ. This project will host in a single VM, but I
> want to use 0MQ just in case I need to scale it up.
> I have drawn a diagram to show my draft plan at:
> http://neo.minidns.net/usbhd/tmp/mcu-dia1.png
> In the beginning(single VM) it will only have one "Core" , one "MCU
> Frontend" and one "Web Frontend". The "Core" is responsible to route
> messages from User to MCU and vice versa. Both the User and MCU can
> initiate to send a message, and there is no need to guarantee a response.
> In case I need to scale up, I can just many Core, Web Frontend or MCU
> Frontend as needed, with minimal configuration change.
> Please let me know if there is significant flaw in my plan or a simpler
> way to do it.
> Thanks and Regards,
> Michael Fung
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev

More information about the zeromq-dev mailing list