Re: Simulation of 400-600 database users

Mercury Interactive's LoadRunner product works pretty good. You basically record the actions that your user does from start to finish. (loadrunner will launch your app, then sit quietly while you do your thing. When
you're finished you click stop on a tool bar). Then you parameterize this script so that you can insert different data into the database and you can tell the software to run 100's of users (again,
loadrunner has an option to help you easily parameterize things). You can set up points where you have users all do the same thing at the same time,
etc. (such as all entering the same record at the same time). So that you can
emulate a peak load. Mercury has had companies run several hundred of users from a single PC.Loadrunner basically does every step excluding the GUI. So you record
the actual SQL statments. Their tool lets you run all of the users scripts that you've recorded
from 1 single place. You can start and stop users, etc. However, to get several hundred users scripts running
from 1 PC you would need to use NT. You should check that they're still supporting
the front end from Windows 95 and not just NT. And yes, I use to work for them. No, I don't
know much about the competitors product. Good luck.

Stephanie

Arnd.Gerns_at_bra.adtranz.de wrote in message <6o7phh$rkr$1_at_nnrp1.dejanews.com>...
>Hi,
>
>we have an application (planning system for public transports), which will
be
>used by 400-600 users. Therefore load tests are an essential part of
testing.
>
>The users will use Windows 95 Clients. Via a wide area network they will
>be connected to a Sun Solaris Oracle Server. The data will be processed
>on the clients, the server will be used only for data management. The wide
>area network isn't expected to be a bottleneck.
>
>Do you know a tool or method, which allows the simulation of 400-600 users?
>It should also be possible to record the actions of some users in order to
>obtain load profiles. These load profiles will then be manifolded during
>the simulation.
>
>Alternatively we could use calculation methods, which allow theoretical
>proofs. Do you know something like this?
>
>Every hint is appreciated.
>
>Thanks and best regards,
>Arnd.
>
>Arnd.Gerns_at_bra.adtranz.de
>
>-----== Posted via Deja News, The Leader in Internet Discussion ==-----
>http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
Received on Mon Jul 13 1998 - 05:41:18 CEST