[zeromq-dev] PGM: multiple listener behavior?

Stuart Levy salevy at illinois.edu
Tue Aug 7 19:33:26 CEST 2012


Thanks - could you explain a bit more?   In the 
PUB+SUB-on-same-port-on-same-host situation Pierre described - which 
applies to me too - *all* NAKs will be lost, not just some small 
fraction of them, if the SUB socket gets opened before PUB, IIUC. Does 
that imply that recovery can never happen, regardless of rate limiting, 
etc.?   I don't want the loss of a single multicast packet to mean that 
the whole system will need to be restarted...

If so, then I don't see how I can use zeromq multicast to create a 
barrier primitive - that I'd be better off just doing all the low-level 
networking myself.  That's a pity, if true.

(There would still be one way to do an N-node barrier within zeromq 
without a central broker: to give each node its own PUB multicast 
address, and have each node SUBscribe to all the others' PUBs.   It 
seems wasteful of multicast space, and needs more config information to 
be passed around, but it would avoid the PUB=SUB port collision. Hmm.)

Thanks for any advice.

     Stuart


On 8/6/12 9:08 PM, Steven McCoy wrote:
> On 6 August 2012 17:34, Stuart Levy <salevy at illinois.edu 
> <mailto:salevy at illinois.edu>> wrote:
>
>     Hmm.  So can we tell what is the consequence of that, at ZMQ level, if
>     NAKs get lost?
>
>
> PGM continues until the protocol breaks then the window will be 
> flushed until it can recover.  A constant high data rate will likely 
> not permit the receiver to rejoin, which is why rate limiting is 
> available.
>
> -- 
> Steve-o

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20120807/e69be5c0/attachment.htm>


More information about the zeromq-dev mailing list