Rodrigo A. Cremaschi wrote:
Hi all,
Some of you asked for the performance tests, so here they are. I tested only SMS section of Kannel (not WAP), using "fakesmsc" on two platforms:
PLATFORM 1: Sun Netra t1, 1 CPU, 400MHz, 512MB RAM, 2 SCSI discs 9GB, Solaris 8. PLATFORM 2: Intel, dual Pentium III, 800MHz, 1GB RAM, 1 SCSI disc 18GB, FreeBSD 4.10.
Tests description: TEST 1: answering text from the configuration file. Kannel sends back a predefined string, 160 characters in length. TEST 2: retrieving a file from the disc. Kannel reads a file and sends the content back. TEST 3: answering from a PHP script. Kannel sends a HTTP request to a PHP script which answers with a 160 character text. The HTTP server runs on a different machine. TEST 4: retrieving data from a database. Same as above, but the PHP reads from a DB. The HTTP + DB server run on a different machine.
Examlpe test: fakesmsc -v 1 -p 10000 -i 0.0001 -m 1000 "1111 2222 text performance" Each test was run 10 times and then averaged.
On PLATFORM 1, TEST 1: fakesmsc takes some processor power to generate the requests and get them back, so it's better to run it on other server (which I didn't). Kannel accepts 1253 messages/s Kannel answers 370 mesg/s TEST 2: Kannel accepts 1143 messages/s Kannel answers 412 mesg/s TEST 3 & 4: not performed.
On PLATFORM 2, TEST 1: Kannel accepts 6675 messages/s Kannel answers 1687 mesg/s TEST 2: Kannel accepts 7366 messages/s Kannel answers 1826 mesg/s TEST 3: Kannel accepts 9945 messages/s Kannel answers 208 mesg/s TEST 4: Kannel accepts 9709 messages/s Kannel answers 86 mesg/s
I will have "real life" tests in a few days.
interesting figures, indeed. Thanks a lot.
I had similiar regression and stress testing while coding our smppbox (Kannel SMPP server). The setup was something like this:
fakesmsc <-> bearerbox(server role) <-> smppbox <-> bearerbox(client role)
with at least an average of 4500-5000 msg/s though the whole chain.
So we'd be pleased to say:
a) it's most unlikely that you will be able to generate real life MO message flows in that ammount
b) in case you would get a) reached, it's most unlikely that your SMSC provider will be able to deliver them in that speed
In short terms: you'll never get the full-speed of Kannel in real life production ;)
Stipe
mailto:stolj_{at}_wapme.de ------------------------------------------------------------------- Wapme Systems AG
Vogelsanger Weg 80 40470 Düsseldorf, NRW, Germany
phone: +49.211.74845.0 fax: +49.211.74845.299
mailto:info_{at}_wapme-systems.de http://www.wapme-systems.de/ -------------------------------------------------------------------
-----BEGIN PGP PUBLIC KEY BLOCK----- Version: GnuPG v1.2.2 (Cygwin)
mIsEP6mcYwEEAMDnUiUwrbb+xwTFWN6TxF2+XZu7/alwJMeCwMBRvXtPZqfjpPhS OkBpU0F4TrVuugz1HINTSaJTYq10AzDQXp5NkyWgckqW79nPAWuOX0dicbJk+cN2 nM2TI4KaxUDe6u8hghNEnH/i2lXsUu9apnP/iixzV81VC2je3uc9hZpnAAYptEVT dGlwZSBUb2xqIChUZWNobm9sb2d5IENlbnRlciAmIFJlc2VhcmNoIExhYikgPHRv bGpAd2FwbWUtc3lzdGVtcy5kZT6ItAQTAQIAHgUCP6mcYwIbAwYLCQgHAwIDFQID AxYCAQIeAQIXgAAKCRABV0w1BqPYRuSqA/wPzsQxao2YePENCtgRTrO86U6zg3sl OcS6CJFI4FZP5h/xD3GRsNH1+MPSvZlomDdpFnr547DGz/Kq9MXuQwVvlVig5yWZ K5dtKp1r5YLhxJQBhfirZbRFFnYmf19f18J8OoS28tuFVftDl1AIwJS3HLyBTv6H g2HyLAEKQIp30Q== =aYCI -----END PGP PUBLIC KEY BLOCK-----
