[Ace-users] Re: [tao-users] performance of COS Notification

[Matthew Gillen]

Friedhelm Wolf wrote:
> 6. I did some measurements and found that even if the latency for 15 
> consumers is different for every consumer (basically it looks like that
> the calls are processed serially and the last consumer has to wait the
> longest to get an answer) the throughput remains the same. I'm not sure
> if I understand it the right way, but should a bad latency not result in 
> reduced throughput?

Not necessarily.  Latency and throughput are often related, but not always.
 In this case, it looks like some consumers were always 'first in line' to
get events, and thus had lower latencies (supposing latency is measured from
publisher-event-generation-time to consumer-event-delivery-time).

What's kind of strange about your results is that the publisher had a
slightly lower publish throughput (410.17 event/sec) than all the consumer's
throughput (410.22 event/sec):

> See the measurements I did below:
> 
> fwolf at simi122:~/SIMSAT/install/cots/ciao/ACE_wrappers/TAO/orbsvcs/tests/Notify/performance-tests/Throughput>
> ./Throughput -ORBInitRef NameService=iiop://localhost:9999/NameService
> -consumers 15 -suppliers 1 -payload 4056 –burst_count 20
> 
> collocated_ec_ 0 ,
> burst_count_ 20,
> burst_pause_ 10000,
> burst_size_  1000,
> payload_size_ 4065,
> consumer_count_ 15, 
> supplier_count_ 1
> 
> Consumer [00] latency   : 147[13149]/289/31184[2592] (min/avg/max)
> Consumer [00] throughput: 410.21 (events/second)
> Consumer [01] latency   : 246[1248]/453/53375[8677] (min/avg/max)
> Consumer [01] throughput: 410.22 (events/second)
> Consumer [02] latency   : 336[19726]/600/73159[8677] (min/avg/max)
> Consumer [02] throughput: 410.22 (events/second)
> Consumer [03] latency   : 428[19726]/737/93101[8677] (min/avg/max)
> Consumer [03] throughput: 410.22 (events/second)
> Consumer [04] latency   : 516[19726]/872/100888[13017] (min/avg/max)
> Consumer [04] throughput: 410.22 (events/second)
> Consumer [05] latency   : 604[19726]/1007/120831[13017] (min/avg/max)
> Consumer [05] throughput: 410.22 (events/second)
> Consumer [06] latency   : 703[19726]/1139/124101[13017] (min/avg/max)
> Consumer [06] throughput: 410.22 (events/second)
> Consumer [07] latency   : 794[19726]/1272/139108[13017] (min/avg/max)
> Consumer [07] throughput: 410.22 (events/second)
> Consumer [08] latency   : 882[19726]/1401/139386[13017] (min/avg/max)
> Consumer [08] throughput: 410.22 (events/second)
> Consumer [09] latency   : 974[19726]/1533/139535[13017] (min/avg/max)
> Consumer [09] throughput: 410.22 (events/second)
> Consumer [10] latency   : 1062[19726]/1665/139654[13017] (min/avg/max)
> Consumer [10] throughput: 410.23 (events/second)
> Consumer [11] latency   : 1150[19726]/1794/139833[13017] (min/avg/max)
> Consumer [11] throughput: 410.23 (events/second)
> Consumer [12] latency   : 1237[19726]/1924/139938[13017] (min/avg/max)
> Consumer [12] throughput: 410.23 (events/second)
> Consumer [13] latency   : 1325[19726]/2053/140051[13017] (min/avg/max)
> Consumer [13] throughput: 410.23 (events/second)
> Consumer [14] latency   : 1412[19726]/2183/140160[13017] (min/avg/max)
> Consumer [14] throughput: 410.23 (events/second) 
> 
> Supplier [00] latency   : 1534[14788]/2425/140278[13017] (min/avg/max)
> Supplier [00] throughput: 410.17 (events/second) 
> 
> Totals:
> Notify_Consumer/totals latency   : 147[13149]/1261/140160[13017]
> (min/avg/max)
> Notify_Consumer/totals throughput: 6153.19 (events/second) 
> 
> Notify_Supplier/totals latency   : 1534[14788]/2425/140278[13017]
> (min/avg/max)
> Notify_Supplier/totals throughput: 410.17 (events/second)
> 
> Any help from people, who have experience with this issues is very much
> appreciated.
> 
> Thanks,
> Friedhelm
> 
> 
> On 7/16/07, *Douglas C. Schmidt* < schmidt at dre.vanderbilt.edu
> <mailto:schmidt at dre.vanderbilt.edu>> wrote:
> 
> 
>     Hi Friedhelm,
> 
>     > TAO VERSION : 1.5.8
>     > ACE VERSION : 5.5.8
>     >
>     > HOST MACHINE and OPERATING SYSTEM:
>     >     i686 pc, SUSE linux Enterprise Server 9, Kernel 2.6.5
>     >
>     > COMPILER NAME AND VERSION (AND PATCHLEVEL):
>     >     gcc 3.3.3 (SuSE Linux)
>     >
>     > THE $ACE_ROOT/ace/config.h FILE:
>     >     #include "config-linux.h"
>     >
>     > THE $ACE_ROOT/include/makeinclude
>     > /platform_macros.GNU FILE :
>     >     include $ACE_ROOT/include/makeinclude/platform- linux.GNU
>     >
>     > CONTENTS OF $ACE_ROOT/bin/MakeProjectCreator/config/default.features
>     >     (used by MPC when you generate your own makefiles):
>     >     N/A
>     >
>     > AREA/CLASS/EXAMPLE AFFECTED:
>     >     COS Notification
>     >
>     > SYNOPSIS:
>     >     performance benchmarks for COS Notification
>     >
>     > DESCRIPTION:
>     > Hi there,
>     >
>     > we are considering COS (RT) Notification as a means to
>     > exchange (a large amount of) asynchronous data.
>     > I guess we are not the first ones out there doing this.
>     > Are there any recent performance benchmarks or experiences from
>     > projects, where the Notification service is used?
>     >
>     > We're especially interested in throughput and scalability
>     > depending on the message size.
> 
>     Please see
> 
>     TAO_ROOT/orbsvcs/tests/Notify/performance-tests
> 
>     Thanks,
> 
>             Doug
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> tao-users mailing list
> tao-users at mail.cse.wustl.edu
> http://mail.cse.wustl.edu/mailman/listinfo/tao-users