[zeromq-dev] What is the exact issue with enabling MULTICAST_LOOP on ZMQ sockets?

[Chinmay Nerurkar]

Marten Feldtmann <itlists <at> schrievkrom.de> writes:

> 
> On 10.06.2013 22:23, Steven McCoy wrote:
> 
> >
> > It is impossible to have multiple reliable publishers on the same host -
> > they cannot both receive NAK unicast messages.
> >
> > Unlike the consumption side the publish side needs broker design attention.
> 
> you mean: "impossible .....same host", when the applications have the 
> same network parameters ?
> 
> To make a reliable publisher one of the following parameters should be 
> different: network adapter, network port, multicast group ?
> 
> Thanks for the clarification ...
> 
> Marten
> 

Hi, Steve,

Thanks for the explanation. We need to have multiple publishers (publishing
to one multicast channel) on the same host. As a follow-up to Marten's
question, would the following design be an optimal solution to the problem?

-A single ZMQ proxy process is started on a host with an incoming XSUB port
bound to "ipc://inbound" and an outgoing XPUB port bound to one multicast
channel "epgm://ethInterface;MulticastIP:MulticastPort".
-All the publishing processes on the host connect to "ipc://inbound" on the
proxy process and publish data.
-The proxy publishes all received data to one multicast channel on its
outbound port. This way there is only one multicast publisher per host.

Also, just to make sure, is it possible to have multiple ZMQ EPGM XPUB
sockets each bound to a different multicast channel, on the same/different
Ethernet adapter on the same host?

Thanks.

Chinmay