Andrew,
looks like I did not express what I mean clear enough.
My point is Credit-Based Flow Control approach suggested by Pieter is not always affordable,
because it involves significant performance loss.
Basically, Credit-Based Flow Control is an external device to analyze gasoline consumption profile,
requires a couple of litres per kilometer.
My suggestion is to utilize some equipment of the car to have this job done with smaller extra.
Guys who care about monitoring are usually performance-focused (like me).
And I think it would be better from performance point of view to have some monitoring facilities inside 0mq.
Pieter,
I'll try to address your request and provide example to explain why it affects performance in a terrible way.
Do hope to have it done until Monday.
Actually it is possible that I'll change my mind creating this example ;)
Thanks.
31.10.2011, 18:31, "Andrew Hume" <[email protected]>:
really?i care about monitoring.and while i have stated here repeatedly that it is a significant pain in the arse thatzeromq won't give me any idea of teh queue length (and please, we udnerstandall about the necessary imprecisions involved -- there is no earthly reasonwhy an estimate couldn't be provided!), i drank the kool-aid and did app-level monitoringand have honestly never looked back.i 'm sure it has a performance impact, but it seems insignificant compared to all the otherwork going on.i do the monitoring in perhaps an odd way, which i could talk about separately,but still flushing messages through a quick recv/send loop seems effcicient to me.andrewOn Oct 31, 2011, at 12:48 AM, Ilja Golshtein wrote:Hello
31.10.2011, 03:24, "Pieter Hintjens" <[email protected]>:* use explicit app-level queues where it makes sense, and monitor those myself
Pieter, people who don't care about monitoring can follow your advice, but they would not
because they don't care.
People who care about monitoring actually cannot,
because suggested approach affects performance in terrible way.
That why I've started this thread.
Regarding getsockopt. Do you believe any attempt to expose peer-specific info
at API layer is not acceptable?
Thanks.
--
Best regards,
Ilja Golshtein.
_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev
------------------Andrew Hume (best -> Telework) +1 623-551-2845AT&T Labs - Research; member of USENIX and LOPSA_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev
--
Best regards,
Ilja Golshtein.
Best regards,
Ilja Golshtein.
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
