Currently i am working on creating Mainframe CICS virtual service to use in Performance testing. Our TPS requirement is 300.
Can you please let me know what is the max TPS will CICS virtual Service supports.
If you have a Functional VSE then througput is 10 tps, independent of transport or data protocol. This is throttled by the product.
If you have a Performance VSE then throughput is only limited by system resources, nbr of CPU's, amount of RAM, amount of RAM assigned to the VSE java process (-Xmx option), network latency, ...
So we cannot tell, but 300 tps is not a extremely hign throughput, so most likely you will be ok. Also, be aware that throughput will be affected by any custom processing you might have added to your vsm like scripting, querying some database, ... In that case even harder to tell.
Hope this helps,
Thank You Danny.
We have Performance VSE and below are our Server configuration:
CPU: 24 Core Processors
Memory: 64 GB
VSE Memory Assigned: 8 GB (in VMoptions file)
And the service is created using Copy book files to parse byte messages. And it is not having any scripting and DB logic.
Could you please give me some idea on how much max TPS can handle with this configuration.
I'm sorry but I would have no idea what the max TPS would be. Your server is adequately configured so with some certainty we could state that 300TPS would be achievable.
What we would do at customer site is not start guessing but actually measure the max TPS for the virtual service that was going to be used in upcoming performance tests. DevTest coordinator/simulator setup has all the features to run proper load testing.
Using some of the requests of our virtual services we would build a few testcases of the most used transactions which we would run against the virtual service. Set up the staging document to run the suite of testcase continuously for a certain period of time with as many concurrent instances (vUsers) as you expect during actual performance testing. We would actually do a few runs and keep increasing the concurrent instances (vUsers) so that we also knew up front at which TPS the virtual service itslef would become the bottleneck, ie. this is the answer to your question 'some idea on how much max TPS can handle with this configuration'