[zeromq-dev] Load distribution vs fail over

[Caroline Leung]

Thanks for the reply.

The failover scenarios that I have in mind are similar to what Gonzalo and
Martin have described in the forum earlier.  
Suggested code snippet copied below:

socket_t s (ctx, ZMQ_DOWNSTREAM);
s.setsockopt (ZMQ_PRIORITY, 1);
s.bind ("tcp://eth0:5555");
s.setsockopt (ZMQ_PRIORITY, 2);
s.bind ("tcp://eth0:5556");

This way ZMQ is acting like a directory service, automatically directing
applications to fail over to connect to the secondary port if the primary is
unavailable.

Thanks you very much for your help!

Cheers,
Caroline


-----Original Message-----
From: Pieter Hintjens [mailto:pieterh at gmail.com] 
Sent: Thursday, April 08, 2010 1:58 PM
To: Caroline Leung
Subject: Re: [zeromq-dev] Load distribution vs fail over

Hi Caroline,

There's no broker, so no need for central failover as you would have
in an AMQP network.  So what kinds of failover are you thinking of?
Recovery from crashed applications?

We're happy to help your company replace its existing middleware with
ZeroMQ.  We would need some details on your application, technical
requirements, and what you expect to win with the move.  I assume its
market data and trades?

Regards
Pieter Hintjens
CEO, iMatix Corp.

On Thu, Apr 8, 2010 at 6:20 AM, Caroline Leung
<caroline.leung at eclipseoptions.com> wrote:
> Hi Zmq developers,
>
> Earlier posts mentioned that the failover capability is on the zmq
> development pipeline, I am just wondering if the team has a tentative
> release date for this feature, as our company is really interested in
using
> zmq to replace our existing middleware layer, but the failover capability
is
> critical in our decision making.
>
>
>
> Cheers,
>
> Caroline Leung
>
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
>
>