Hello,

We have a JTAPI application, that features: monitoring stations and invoking listenHold/listenUnhold for calls on demand. During load testing our application we discovered that, even though there are no issues with the application until 15-20 sessions (simultaneous calls with selective listen hold), further listenHold invocations are refused with CSTA Error: 31 (genericSystemResourceAvailability).

I've built a sample application to test the "throughput" for listenHold only in the same lab environment. There are 100 simultaneous calls on which we invoke listenHold. The first 75 is executed, the last 25 is refused with the same CSTA Error: 31.

Can you suggest where could be the bottleneck or how to identify the limit that we encounter?

ecsjtapia 7.1.0.8
AES: 7.1.3.3
CM: 7.1.3.0.0-FP3

Thanks, Sandor

The main bottlenecks between the application and AE Services are:

1. The application sends too many requests or receives too many events per second. According to the AE Services overview document, AE Services can handle 2000 msgs/sec so this seems unlikely to be your problem here.

2. The application does not quickly release the SDKs listener threads. This can lead to a backlog of events/responses in the input buffer and cause problems.

However, according to the TSAPI programmers guide:

GENERIC_SYSTEM_RESOURCE_AVAILABILITY (31) (CS3/40) = Switch capacity has been exceeded.

This seems to imply that the problem is due to a limitation on Communication Manager/ System manager or the media server/media gateway.

My suspicion is that SLH is causing you to use extra media resources on your media gateway/server. The SAT command 'list media-servers' or "list media-gateways" may let you know if you have used up all your media resources.

Martin