Application Development Discussions
Join the discussions or start your own on all things application development, including tools and APIs, programming models, and keeping your skills sharp.
cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 

Acceptable Response Times for a Development(ST05 & SE30)

Former Member

‎10-11-2008 6:52 AM

0 Kudos

Hello Everybody,

Can anybody tell me a way on how to find the acceptable resonpse times for a development in ECC. I mean, when we check the Runtime Analysis, we get the DB and ABAP times. As the data is very less in terms of business now, the reports or the developments are working fine, but tomorrow, the data becomes large these will have the performance issue.

So, based on this can we generalize and find out the acceptable percentages in SE30 as well as the times of execution in ST05. And can we combine the outputs of these two and come to a conclusion on the performance of the development.

Regards,

Sreenivas

2 REPLIES 2

Former Member

‎10-13-2008 8:13 AM

0 Kudos

Hi,

SE30:

what you call for is benchmarking:

  • small-time benchmarking:

compare several solutions to a problem to find the fastest one

this is benchmarked in a (mostly) single user system

  • big-time benchmarking:

increase the data volume in several steps

increase the amount of user's in several steps

Measure the runtime and draw a graph that can describe the performance you encounter and

extrapolates that for different data volumes and users.

For simulating user's you can use some tools availbale for a stress test.

ST05:

For ST05 I use an export of the database statistics from the productive system and import them into development for the relevant tables. So I don't need the REAL data in the dev tables but can generate execution plans that will reflect the plans later on the prod. system.

bye

yk

former_member194613
Active Contributor

‎10-13-2008 8:18 AM

0 Kudos

> but tomorrow, the data becomes large these will have the performance issue

what do you mean with tomorrow

Tomorrow, Tues Nov 14 or just future?

If the first one, then you are probably too late to fix something.

If the second, the you should be able to extrapolate from measurements with

10 objects to 100 and 1000 and 10.000.

Buut be careful besides the linear extrapolation you must take care that there is no nonlinear behavior,

Nonlinearity: The problem and background

Z_SE30_COMPARE

Nonlinearity Check

Siegfried