Load Testing Results : 120 Lines (60 Out And 60 In

[SupportTeam]

Here are some results of a load test we recently did.

 

System specs:

 

P4 3GHz, 1GB RAM, SATA HD

WinXP SP2

SR6.0 SU84

DMV1200B card.

VoiceGuide for Dialogic v6.0.2430

 

The DMV1200B card has 4 E1 ports. These ports were looped back to each other so calls going out on first E1 line would arrive on the 3rd, and calls made on 2nd E1 line would arrive on the 4th.

 

VoiceGuide was started and about 10,000 calls were loaded using the telephone numeber loader. Outgoing calls were set to run script "cc test 1.vgs" upon connection. Channels accepting calls were setup to run the "Credit Card Payment" demo script.

 

The "Credit card Payment" demo script was designed to log relevant info at the end of the call into the file output.txt.

 

The outgoing "cc test 1.vgs" script was a series of modules which would generate the DTMF tones or play sound files

 

The scripts used, the output, basic tracing and the CPU usage during the test are available in the attachment provided.

 

Basically it shows that when running intensive scripts on all 120 lines the CPU usage was reasonable (30% average) and that the memory and handles usage did not grow during the calls made.

 

The output.txt file also shows that no DTMF tones were 'missed' by the system.

 

The DTMF tones sent over the system have been generated at a much faster rate then what a human caller would press (see sound files used), resulting in the lines being tested at a much more active level then what they would experience on average.

 

Logging was reduced to a minimum level (1).

LoadTestResults.zip

[Guest mhsamei]

Great

but can you test a script with about 400 scripts and 100 run vbs acript on this server?

[SupportTeam]

If you want to post the VoiceGuide script and sound files to use for testing we can set up the test with your script/wavs and report the results here (including logs, screenshots, etc).

 

Would not anticipate any issues - sounds like using your scripts the load on VoiceGuide will be less then in the test described beforehand as most of the time VoiceGuide would probably spend waiting for the VBScripts to complete... (VBScripts are launched as a separate process)