[zeromq-dev] Using 0MQ with an existing messaging library

[Pieter Hintjens]

On Thu, Oct 6, 2011 at 10:56 AM, asif saeed <asif.lse2 at gmail.com> wrote:

> Would you elaborate please? From what I have gleaned from 0MQ docs is that
> it kinda provides its own socket level abstractions. MyLib uses direct
> Windows sockets (async sockets / IO Completion Ports). Windows IO Completion
> works directly with native sockets. I don't immediately see how 0MQ will fit
> in here. Would you like to elaborate a bit more on it please?

Lacking information on your use case, this is going to be a bit vague.
Imagine you want to use both HTTP and 0MQ. The rationale is that you
have existing HTTP applications (web browsers) and you want to connect
them to new 0MQ apps. You create an app that plugs into a web server
and maps certain HTTP requests like PUT/GET to 0MQ send/recv commands.
This is a bridge. You can then use that bridge to create networks of
networks, i.e. a 0MQ network connected to a HTTP network connected to
another 0MQ network.

This is a standard way to mix two transports. You may be able to do
the same with 0MQ and MyLib. It does depend a lot on your
requirements, which you haven't really explained.

-Pieter