How much Asterisk server can handle simultaneity calls

Asterisk server:
2gb Ram
CPU duocore 2660/1066/4M

will make outgoing sip calls to Siemens Hipath8000

How much approximately such Asterisk server will can handle simultaneity calls…

For any answer!

It depends on a few factors. What are you doing with asterisk? Bridging the calls and re-inviting out? Or doing MeetMe, Call Recordings etc… There is a recent thread on the asterisk-users list regarding this: … er/209208/

If you want close to a definitive answer. Run a sipp test against your dialplan on your hardware to see at what point things start to error. For a simple comparison, On my dev server (Roughly 5 year old single 2.4 GHz), I’ve run a sipp test sending the calls into Wait(10) then hangup. I was able to get over 900 concurrent channels up without an issue. But, if I changed it to Playback() or Background() a message, I was only able to get 140 or so channels up before things started erroring. So it largely depends on what you are doing.

I’ve a server that went as high as 450 SIP Calls once, ULAW, No transcoding, RTP stream going through. (No SIP re-invites), no realtime or caching, this user is hardcoded in the configuration file.

Dual Intel Xeon 2.8
2 gigs of ram
Asterisk 1.4.17

Right now it is at:
320 active channels
162 active calls

10:45:25 up 108 days, 17:22, 2 users, load average: 0.60, 0.54, 0.39

free -m:
total used free shared buffers cached
Mem: 2026 1144 881 0 216 772
-/+ buffers/cache: 155 1870
Swap: 1027 0 1027


you mean you did not reboot in 108 days?
in some threads, they recommend to reboot every day…

This is very comforting because I began to get scared to put my users with a system that have to reboot every day…

Which OS are you using on the host?


Asterisk PBX uptimes are usually measured in months and years. To compete with the reliability of the PSTN, ensuring a stable system is paramount. CentOS (i.e. Red Hat) is the most popular Linux flavor to run Asterisk on.

You shouldn’t have to reboot at all unless you upgrade the kernel. There have been significant amount of bug fixes to 1.4 since its release. Any issues you’d have now most likely would be edge cases.

Here are the uptimes from our main office pbx. It is idle right now.

06:24:11 up 195 days, 18:04, 1 user, load average: 0.00, 0.01, 0.00

System uptime: 27 weeks, 6 days, 18 hours, 3 minutes, 40 seconds

Actually I was showing you the uptime of the machine to show you the actual CPU Load of the machine with 160 active calls.

About Asterisk…

ca1*CLI> show uptime
System uptime: 5 weeks, 21 hours, 38 minutes, 30 seconds

It never need to be restarted if your install is stable. The only time I restart is when I recompile something in Asterisk and that the module is not a module that can be reloaded via the CLI.

The OS i use the most at the moment is CentOS 5, but i’ve old installs on servers that have been up and running for almost 2 years now using Asterisk 1.2 running of some old version of fedora. As long as you keep the daemons running ports open to the internet up to date with yum and keep your asterisk branch updated to the latest version, i’ve had no security issues.

I run 8 medium-heavy load asterisk (7 using Realtime/MySQL peers and cdr) so if you have any questions don’t hesitate.