Performance testing. How much data should I create?

Question

I am very new to the development of technical specifications, so I have a very simple question.

I work on a client-server system using the SQL server backend. The application is a huge tax-related application that requires performance testing at maximum load. This means that the system should have 10 million tax returns when we run the scenarios associated with the creation of tax returns and their sending. Then there will also be a proportional number of users to be created.

Now I hear in meetings that we need to create 10 million records to test performance and run scripts with 5,000 users, and I just don't think it's possible.

When we talk about creating a smaller dataset and extrapolating performance planning, the very general answer I hear is that we need 10 million records because we cannot tell from a smaller dataset how the database will behave or network.

So, how can you plan performance and test performance on a large enterprise application without creating a peak data level or running the maximum number of scenarios?

Thank.

Answer 1

Now I hear in meetings that we need to create 10 million records to test performance and run scripts with 5,000 users, and I just don't think it's possible.

Why do you think so?

, ( ) , , ( ) , .

: , ; / , , . , ( ). SQL , .

, - , , ( ) .

, , , , . , , : -).

:

, ( " " ), . , : -).

Answer 2

, . , " ". , . , , 10 .

, 10 , 11 . , , 100 , . , "10 "

Answer 3

Redgate SQL Data Generator. .

Answer 4

" /Ian Molyneaux, OReilly, 2009".

Answer 5

. , . ApacheBench, .

Answer 6

, ruby ​​faker perl data faker. . SQL redgate .