[zeromq-dev] Hi All

Michel Pelletier pelletier.michel at gmail.com
Mon Dec 2 18:54:55 CET 2013

The two cases you are comparing are not equivalent.   The socket example
simple streams data from one process to another, but REQ/REP are
synchronous, you are issuing a "fetch" request and waiting for the reply.
 That round trip takes time that your socket example does not replicate.

Read the guide, use PUSH and PULL sockets, and loose the whole "fetch"
thing and you'll get the same performance as with sockets.


On Mon, Dec 2, 2013 at 9:49 AM, Bhanu Chouhan <
bhanupratapsinghchouhan88 at gmail.com> wrote:

> Hi All,
> I have tried some request and reply example.
> I have written a simple server and client in BSD sockets. In which server
> (file_bsd_ser.py) binary read a file requested from client
> (file_bsd_req.py) and send it to client in chunks.
> I have done the same using ZMQ REQ & REP sockets in filerepexp.py (server)
> and filereqexp.py (client).
> The file i have chosen was of 50 MB.
> The task is taking almost double time in case of ZMQ, but it should not be
> the case.
> Please please please help me on above query that why ZMQ taking more time.
> Actually it is required for my project on ZMQ study.
> regards,
> Bhanu
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20131202/5ee649ca/attachment.htm>

More information about the zeromq-dev mailing list