On Jul 18, 2013, at 7:07 PM, Steven McCoy wrote: > On 18 July 2013 16:12, Diego Duclos <[email protected]> wrote: >> I would have use for some of this stuff as well. >> However, I don't think it really feels at home in zmq itself ? What about a >> middleware layer between zmq and user apps that takes care of this ? > > > There is a cost on the send side, you would probably want to implement a > watermark a factor of the network latency and desired lifetime and still gain > some benefit from network batching. > > It is technically feasible with PGM to kill the messages in the transmit > window such that NAKs will be denied and late joiners will not ask for them.
I second Steve; this is something that is best inside of ZMQ as ZMQ really knows when it is about to send a message out. Killing the messages in the transmit window is EXACTLY what I was thinking about! Steve, would this be easy to implement in OpenPGM? I'm willing to look at this, but since I don't know the codebase, I'm not sure where to start. Thanks, Cem Karan
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
