[zeromq-dev] Asynchronous Client-Server Pattern in ZGuide

Pieter Hintjens pieterh at gmail.com
Thu Nov 15 08:09:05 CET 2012


I don't think any of the designs would put a router at the worker side, it
is an option at the server side.
On Nov 15, 2012 3:51 PM, "Ilja Golshtein" <ilejncs at narod.ru> wrote:

> Yes,
>
> so the statement "This would require that workers start by telling the
> server they exist"
> is incorrect because we have Router at workers' side.
>
> 15.11.2012, 10:37, "Pieter Hintjens" <ph at imatix.com>:
> > If you use a DEALER/ROUTER connect then the dealer has to connect to
> > the router before the router can reply to it. That's the
> > load-balancing pattern already explained.
> >
> > -Pieter
> >
> > On Thu, Nov 15, 2012 at 3:35 PM, Ilja Golshtein <ilejncs at narod.ru>
> wrote:
> >
> >>  Pieter,
> >>
> >>  my question is if the description of
> >>  what should be changed to switch
> >>  from DEALER/DEALER to DEALER/ROUTER
> >>  correct.
> >>
> >>  15.11.2012, 03:20, "Pieter Hintjens" <ph at imatix.com>:
> >>>  Ilja,
> >>>
> >>>  Correction, the text is right, though perhaps confusing. It explains
> >>>  two options (router-dealer, dealer-dealer), and then explains why we
> >>>  chose dealer-dealer in this case.
> >>>
> >>>  -Pieter
> >>>
> >>>  On Thu, Nov 15, 2012 at 8:17 AM, Pieter Hintjens <ph at imatix.com>
> wrote:
> >>>>   You're right, the text is wrong. The server and workers talk using
> >>>>   DEALER-to-DEALER. I'll fix the text, thanks for catching this.
> >>>>
> >>>>   -Pieter
> >>>>
> >>>>   On Thu, Nov 15, 2012 at 1:10 AM, Ilja Golshtein <ilejncs at narod.ru>
> wrote:
> >>>>>   Hello,
> >>>>>   Pieter,
> >>>>>   All.
> >>>>>
> >>>>>   In ZGuide in "The Asynchronous Client-Server Pattern" section we
> have
> >>>>>   ==
> >>>>>   Workers get unaddressed messages, and we manage the connections
> from server thread to worker threads explicitly using a ROUTER socket as
> backend. This would require that workers start by telling the server they
> exist, which can then route requests to workers and track which client is
> 'connected' to which worker. This is the load-balancing pattern again.
> >>>>>   ==
> >>>>>
> >>>>>   Since we are talking about Routers at workers' side, I am not sure
> why "This would require that workers start by telling the server they
> exist".
> >>>>>   It is server who should start a communication and he does it
> anyway.
> >>>>>
> >>>>>   Actually switching from DEALER to ROUTER at worker's side requires
> reading and writing extra frame, which is identity.
> >>>>>
> >>>>>   Do I miss something?
> >>>>>
> >>>>>   Thanks.
> >>>>>
> >>>>>   --
> >>>>>   Best regards
> >>>>>   Ilja Golshtein
> >>>>>   _______________________________________________
> >>>>>   zeromq-dev mailing list
> >>>>>   zeromq-dev at lists.zeromq.org
> >>>>>   http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> >>>  _______________________________________________
> >>>  zeromq-dev mailing list
> >>>  zeromq-dev at lists.zeromq.org
> >>>  http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> >>  --
> >>  Best regards
> >>  Ilja Golshtein
> >>  _______________________________________________
> >>  zeromq-dev mailing list
> >>  zeromq-dev at lists.zeromq.org
> >>  http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> >
> > _______________________________________________
> > zeromq-dev mailing list
> > zeromq-dev at lists.zeromq.org
> > http://lists.zeromq.org/mailman/listinfo/zeromq-dev
>
> --
> Best regards
> Ilja Golshtein
> _______________________________________________
> 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/20121115/2c721e0c/attachment.htm>


More information about the zeromq-dev mailing list