vicidial.org

Hi guy’s,
This is my build:
ViciBox v.7.0.4-170113
SVN Version : 2656
DB Schema Version: 1484
Sever: Express
Version: 2.14-583a
Build: 161226-2224

I’m currently using a:
Dell Poweredge 2950;
Intel(R) Xeon(TM) CPU 3.00GHz (2x);;
PERC 5/i;
2x 146GB 15k mirror;
4GB Memory;

I’ve got around 25 agents at the moment. There all connected to the server with hardware phones. I’ve got Grandstream 1625, Yealinks T20, Aastra 9143i and Aastra 9480i. At the moment the agent are not using the web interface to call customers. They are all calling manually. When they call the customers, we are facing some problems. Some have echo, some have noisy calls, some have latency I takes 30 seconds that the customer answer, like a delay. That’s the first.

When I’m using the software, customers ear us bad. They feel like we are far away or like we were speaking in water.

I’ve been looking into server performance, and I didn’t found any performance issue.

In overall, all I’ve got is bad quality calls. I would like to know if my setup is bad.
I use G729 codec from asterisk.hosting.lv
I tried to switch to G711 (ulaw) , but it went more choppy.
I don’t use the recording function, I think, I didn’t enabled it.
I’ve monitored my network to see if I’ve a bandwith issue.
On the wan I’m always below 4mbps.
I’ve a 30/30 internet connection.
I’m using HP 2920 and 2910 switches.
And for the router I’m using a Cisco RV130W.

I’m open to all suggestion, I just wanna make it work.
Thank you.

Intel(R) Xeon(TM) CPU 3.00GHz (2x);;

Vague. This could be 2x2 (4 cores) or 2x4 (8 cores). Dropping in the CPU ID is not a bad idea, but the core count is essential. No "htop" is not a valid method to find out, depending on the hyperthreading settings you could get the same core count with either setup.

Onward:

Have you changed carriers? (Having three carriers is a must if you intend to have a paycheck for over 25 people riding on this system).

Do all your agents experience the same call quality issues (not from memory here, actually check this personally). Call quality issues are often caused by cheap audio hardware and can even be caused by having too much local traffic or bad network switches.

You did not mention the total number of live calls on your server during these issues, or the Average Server Load. These can both be helpful in troubleshooting audio issues.

If your call quality sucks and you change to ulaw and it Gets Worse, I tend to lean toward bad internet. 4M is kinda small. You should consider outlawing ALL internet usage in the office except for Vicidial (when it's running). And consider running with 50% load for an hour (half the agents, don't increase the dial ratio) to see if the problems are gone. It's not a solution, but it is a good troubleshooting tool. We've had a few call centers that add a second "surfing" ISP so agents can hit the net without interfering with the Vicidial audio.