On Sat, Oct 26, 2013 at 3:59 PM, Arnaud Loonsrta <[email protected]> wrote:
> It would be the same as is happening now. Which is fine but I'd rather > have a system which at least tries to prevent this. I know it's not a > reliable option especially under network stress but it's just a single > packet which could prevent a lot of other packets. > > It could be done as well using unicast methods within ZRE but it's much > more work to send a packet to all nodes. A single broadcast/multicast is > cheaper and if it doesn't work there is still the timeout method. Do you have a simple example case we could work through? Preferably something real, not theoretical. -Pieter _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
