Re: Looking for Workload and Capacity Models, software, docs or tempaltes.

From: Randy Case <randy_case_at_nkn.net>
Date: 1996/05/15
Message-ID: <randy_case-1505962346090001_at_news.nkn.net>#1/1


In article <4n6edc$5jp_at_kettle.magna.com.au>, dedata_at_magna.com.au wrote:

> I work for a computer consulting company in Australia. We have a project that
> requires us to produce a workload and capacity model (as an excel
 spreadsheet)
> for a proposed client/server system.
>
> I was wondering anyone has any tools,templates or documentation that may
> assist us to meet a very tight deadline. Otherwise we will need to build the
> models from scratch.
>
> Any assistance provided will be most welcome. Thankyou.
>
> Regards
> Frank De Palo
> De Data Pty Ltd
> dedata_at_magna.com.au

Been there, done that.

What you have (and / or need) is a "day in the life" of the user. To describe the throughput of the system, modeled as a series of response functions, you need to determine what steps a user will perform, how long each takes, what inputs/outputs from the system result, etc.

This derives two parts of the problem - the network connections (and needed capacity) and the work load on the server(s). The third part is the estimate
of the processing needed to support each function.

You then "stack" each of the scenarios together; i.e., start each scenario at time zero. You also need so scale each scenario by the number of users assumed for each function.

Lastly, I would recommend that you use Monte Carlo on top of Excel to account for variability in the number of users, their response time, size of the data blocks transfered, etc. I have used Crystal Ball for this (_at_Risk will also work).

E-Mail me if you have additional questions at either my work or home address (or both since we are about 15 hours apart in time zone...)

Randy

-- 
Randy Case               |#include <std_disclaimer.h>  |Ask me about
randy_case_at_nkn.net (home)| They don't speak for me,    | Systems
rcase_at_esy.com      (work)|   and I don't speak for them|  Engineering
-----BEGIN GEEK CODE BLOCK-----
       Version: 3.1
       GCS/E/IT/M/S d>+ s: a+ C$ UBVISC*++$ P+ L- !E-
       W+++$>$ N++ !o K-? !w-- !O- M++$ V PS+ PE+_at_ !Y PGP !t
       !5 !X !R- tv? b+++ DI++++ !D G e++ h---- r+++ y++++ 
------END GEEK CODE BLOCK------ 
Received on Wed May 15 1996 - 00:00:00 CEST

Original text of this message