[zeromq-dev] 0MQ protocol stack

[Andrew Hume]

i don't want to appear persnickety about terminology, but there
is a real difference between hop-heartbeating, which i might normally call
keepalives, and end-to-end heartbeats, which really do reflect application logic.

the problem with trying to put reliability/resiliency into 0mq is that its a fair bit of work,
most of which has nothing to do with 0mq per se. in my last project, we had a distributed
hash table thing for modest intra-cluster data sharing, and even on a reasonably
well maintained LAN, it was amazing to see the various error states we reached.

on the other hand, i could see resilient versions of some of the common patterns being
developed as separate projects (like zapi). that may be the best way to approach this.

it would be great if this were an orthogonal thing that could just be added on top of
your favourite topology, but i suspect its not.

	andrew

On Apr 8, 2011, at 5:03 AM, Paul Colomiets wrote:
> 
> Heartbeating and connection init are inherently hop-to-hop, So
> whats wrong with that?


------------------
Andrew Hume  (best -> Telework) +1 623-551-2845
andrew at research.att.com  (Work) +1 973-236-2014
AT&T Labs - Research; member of USENIX and LOPSA




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20110408/a3526633/attachment.htm>