Hello,

we are currently faced to the requirement to monitor and record a few hundreds of extensions in parallel with our DMCC call recording solution.

Basically there are two ways we can go: Service Observing or Single Step Conference.

Do you have any practical experience which of both methods is the better one in such a heavy load situation?

Are there any differences between SSC and SO regarding:
- generated load on CM or AES ?
- delay for establishing the internal conference ?
- consumption of internal resources on CM and AES ?
- ...

Regards,
Claus

Claus,
Have you reviewed the content of https://devconnect.avaya.com/public/flink.do?f=/public/download/AES/WhitePapers/AEServicesCallRecording_SVC4050.pdf

I think you will find some relevent information therein, including a third method of doing call recording.

- generated load on CM or AES ? >>> yes
- delay for establishing the internal conference ? >>> yes
- consumption of internal resources on CM and AES ? >>> no (other than traffic/load differences)... same hardware resources required for either model.

John,
First of all thankyou for providing us with the document.

After reading through the paper, some things are clearer than before, but we still have some questions.

1)
Is there a difference in time the CM needs internally to establish SO in comparison to set up a SSC (not taking in account the time our app needs to process the ESTABLISHED event an to send out the SSC command).

2)
How many SO's or SSC's can a CM handle in a fixed period of time. Is this limitation equal for both types?

The 'Multiple Registrations' method is not a choice for us, because we must also be able to record SIP devices.

Regards,
Claus

1) Yes, there is a difference. SO is less internal work/messaging. Can you tell? Probably not.

2) what do you mean by "in a fixed period of time"? transactions per second? Or system limits?

Hi John,

We need to find out how many SO and how many SSC requests a CM can handle in a given timeslot, e.g. a second.
Especially it is important for us to know, iif the limitations of both are lying in the same range, or if a CM is able to handle noticeable more SOs than SSCs.

Regards,
Claus

Given SO is done purely in Communication Manager and there are some events that are sent to monitors of the Call/VDN/Stations the limitation will be performance bound on CM.

To trigger SSC you need to be able to send a message into Communication Manager through CTI. Thus the message limit per second through the switch connection is a potential bottleneck.

That tied to the increased number of messages and the context switching needed to achieve SSC would lead to better performance from a SO solution.