On 20 March 2011 16:22, Pieter Hintjens <[email protected]> wrote:

> On Sun, Mar 20, 2011 at 3:09 PM, Marko Mikulicic
> <[email protected]> wrote:
>
> > So the focus is on using HTTP as a transport protocol, between zeromq
> peers,
> > not intended to be used as an interoperability solution, did I understand
> > correctly?
>
> Yes to the first part, but for me that precisely means

"interoperability", if the mapping is documented and reliable.
>

Sorry, I mean direct interoperability, as if the unmodified HTTP clients
were able to connect to http/0mq endpoints. For example imagine a SUB
endpoint listening on a http socket;
a http client like a browser or wget or whatever then connects and POSTs a
file which then can be interpreted by the SUB socket as a valid pubsub
message.

On the other hand, a simple HTTP transport encapsulation working as a tunnel
of the 0mq protocol, would not allow this scenario; both the client and the
server have talk the underlying 0mq protocol, with bidirectional
communication. I didn't dig in the 0mq protocol internals but I guess there
is some kind of handshake to handle the identities (perhaps it's optional?).

Cheers,
Marko
_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

Reply via email to