On Mon, Jul 12, 2010 at 1:45 PM, Chuck Remes <[email protected]> wrote: > On Jul 12, 2010, at 12:18 PM, Steven Clark wrote: > >> Forgive me a simple/stupid question here- >> >> To make use of ZeroMQ, do both ends of a simple network connection >> need to be programmed with ZeroMQ, or can one endpoint use ZeroMQ and >> the other use low-level (Berkeley-style) sockets? >> >> I have a situation where a device (microcontroller-based) connects >> periodically to a TCP server, sends/recvs some data, and disconnects. >> The server sends some filtered form of the data to other network >> endpoints. Currently everything is written using low-level sockets. >> I'd like to switch the server to use ZeroMQ so I can take advantage of >> Publisher/Subscriber sockets, etc for the distribution of data from >> the server. Should/can I make the link between the device and the >> server use ZeroMQ, given that I can't exactly port ZeroMQ to run on >> the microcontroller platform? >> >> In the event that I can use it for the device connection, is the PAIR >> topology the correct one to use? What happens if the device doesn't >> shut down the connection cleanly (which happens fairly often)? Is >> there any danger of getting stuck in a FIN_WAIT state? > > You can use regular sockets but you need to conform to the 0mq wire protocol. > > See: http://api.zeromq.org/zmq_tcp.html > > I think you could probably get something working on the microcontroller > platform for SUB sockets pretty easily. >
Correct me if I'm wrong here.. zmq_poll() can also be used. And I believe Martin is working on making this transparent for the future. ~peter > cr > > _______________________________________________ > zeromq-dev mailing list > [email protected] > http://lists.zeromq.org/mailman/listinfo/zeromq-dev > _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
