On 09/27/2010 03:45 PM, Mikael Helbo Kjær wrote: > Maybe a bit redundant, but I have to agree. > > With OpenPGM 5 arriving for zmq, ipc support on Windows is the last > major difference in transports between the platforms. Each difference > does reduce the usefulness of zeromq for me as I have clients running > on windows and linux talking to servers running on windows and linux. > Not having caveats in my software design for ipc on windows would > help me.
Ack. Using named pipes requires some polling mechanism that works on named pipe HANDLE first though. > In fact that would leave me with only two wishlist items > (not that I am greedy I can make do most excellently with what we > have now) : > > - reenable of IPv6 on all relevant transports (and maybe the OpenPGM > 5 stuff helps there too), which is off according to my understanding > of the issues list in git Yes. What's missing here is someone with IPv6 expertise to define exact semantics for IPv4/IPv6 interaction. The reason why original IPv6 support was disabled was that it broke IPv4 :( > - some sort of encryption support (I am going to try to tunnel zmq > through SSL because our data might need it, I might try and make some > sort of SSL transport one day). My understanding is that encryption support should be layered on top of 0MQ. The rationale is that if the whole traffic is encrypted even intermediary nodes (devices) would have to be able to decrypt the messages to get the routing info (topics, identities, etc.) This in turn may be a problem in scenarios where intermediary node isn't trusted. Martin _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
