loadrunner How do i design the below scenario ?? 100user...

[Akash Mohanty]

How do i design the below scenario ?? 100user and 10 hours time and 500000 transactions ,what should be the pacing time for each iteration . Also share the calculation trick

 

[Sagar Aggarwal]

Use little's law

 

[Akash Mohanty]

According to my calculation it should 7.2 sec.Just cross check with u ppl.

 

[Roylen Topno]

Share your calculation

 

[Akash Mohanty]

Can any one suggest me the right path to calculate this scenario??? or You could share your Calculation technique ..

 

[Vijay Kumar Reddy]

Pacing=Nf Users*Total duration of the test/Total Nf Transactions.According to that formula Pacing P=100*3600*10/500000=7.2.

 

[Sai Rama]

Is One scenario contain many scripts or one script?

 

[Sasi Anand]

U can use this formula.. User= (pacing*Targetvolume)/3600 which is used to calculate one hour target volume. For the above scenario use user=(pacing *Targetvolume)/(3600*10)

 

[RoHan KoKate]

First calculate transactions per sec... I.e how many transactions will be performed in a peak hour... Let's say 36k in an hr...then divide by 3600 to get TPS....so now the TPS is 10....divide by no of users which is 100...so pacing will be 10 sec that is fixed...you may use random pacing which can be set at 9 and 11..I hope this is what you looking foe

 

[Saravanan Kanagarethinam]

You can check on Goal oriented scenario design on controller/PC as well.

 

[James Pulley]

Can you please define what is a "transaction" in your case. There are roughly 8 different definitions from hit, progressing through page to session and onto completion of a business process. Until you know what your transaction definition represents you are dead in the water. But, assuming uniform distribution after rampup 500,000 in 10 hours = 50,000 in one hour 50,0000 into 100 users equals 500 into one user 500 for one user in an hour (assuming uniform distribution) equals one every 7 seconds (and change) But, if your transaction is a hit, and you get dozens of hits per page loaded, then your model has to be rethought. 100 also seems rather arbitrary a number as does 500,000. I would look to tie both to objective evidence in logs from production