We rejected a straight SCTP transport in favor of ØMQ, primarily because there was no certainty that it would interoperate and perform across vendor solutions. ØMQ offered what we needed without the complexity, over well-understood protocols.
On Oct 27, 2010, at 5:41 PM, Martin Sustrik wrote: > On 10/27/2010 10:03 PM, Chuck Remes wrote: > >> SCTP is an actual transport protocol. >> >> 0mq is a socket library that can be layered on top of any kind of >> transport. Right now it supports "inproc" (comm between threads), >> "ipc" (interprocess communication), "pgm" (multicast), and "tcp" (you >> know what this is). >> >> No one has done this yet, but you could create a new transport called >> "sctp" so that 0mq could support that too. > > Well, once there was an sctp transport for 0MQ but, unfortunately, there > was not much interest in it, so it finally got dropped. > > I, personally, believe that sctp+0mq would be a great combination. While > there's a little overlap between the two (both do transport messages > instead of raw bytes) most of the features nicely complement each another: > > SCTP: multihoming, native L4 heartbeats, multiplexing without need to > establish new connections etc. > > 0MQ: messaging patterns, seemless interoperation with different > transport such as inproc, language bindings etc. > > Martin > _______________________________________________ > 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
