[prev in list] [next in list] [prev in thread] [next in thread]
List: grinder-use
Subject: Re: [Grinder-use] Real world scenarios
From: Matt <mattmoran76 () gmail ! com>
Date: 2008-04-17 13:23:11
Message-ID: 68fea9390804170623i3cb77a9djc8e2737cc9fcbcf0 () mail ! gmail ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
Best way to do this IMHO is to profile the user. In other words, get a
'real' user to use the app and profile the usage.
As i'm sure you know, it's all about the time where the client does nothing,
profiling a user going through a typical case will give you this 'average'
figure.
>From there you can load your 1000 users but with only n being simultaneous.
I like using a percentage of the target user base for throughput testing,
but never for 'real' usage testing. It all depends on what type of app for
sure.
On 17/04/2008, john morony <jm1710@yahoo.com> wrote:
>
> Does anyone have any metrics for calculating expected concurrency for a
> web application given a total number of users in a user base?
>
> For instance, if we expect our application to have 1000 users total,
> what is a good number for expected number of users using our system at
> the same time? 10%, 20%?
>
> My problem is that we can devise tests where our app performs
> beautifully; we can also design tests where we can bring our app to a
> grinding halt. How do we know which one is the more accurate portrayal
> of real life?
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> grinder-use mailing list
> grinder-use@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/grinder-use
>
[Attachment #5 (text/html)]
Best way to do this IMHO is to profile the user. In other words, get a \
'real' user to use the app and profile the usage.<br><br>As i'm sure you \
know, it's all about the time where the client does nothing, profiling a user \
going through a typical case will give you this 'average' figure.<br> \
<br>From there you can load your 1000 users but with only n being \
simultaneous.<br><br>I like using a percentage of the target user base for throughput \
testing, but never for 'real' usage testing. It all depends on what \
type of app for sure.<br> <br><div><span class="gmail_quote">On 17/04/2008, <b \
class="gmail_sendername">john morony</b> <<a \
href="mailto:jm1710@yahoo.com">jm1710@yahoo.com</a>> wrote:</span><blockquote \
class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt \
0pt 0.8ex; padding-left: 1ex;"> Does anyone have any metrics for calculating expected \
concurrency for a<br> web application given a total number of users in a user \
base?<br> <br> For instance, if we expect our application to have 1000 users \
total,<br> what is a good number for expected number of users using our system at<br> \
the same time? 10%, 20%?<br> <br> My problem is that we can devise tests \
where our app performs<br> beautifully; we can also design tests where we can bring \
our app to a<br> grinding halt. How do we know which one is the more \
accurate portrayal<br> of real life?<br> <br> <br> \
-------------------------------------------------------------------------<br> This \
SF.net email is sponsored by the 2008 JavaOne(SM) Conference<br> Don't miss this \
year's exciting event. There's still time to save $100.<br> Use priority \
code J8TL2D2.<br> <a \
href="http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone"> \
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone</a><br> \
_______________________________________________<br> grinder-use mailing list<br> <a \
href="mailto:grinder-use@lists.sourceforge.net">grinder-use@lists.sourceforge.net</a><br> \
<a href="https://lists.sourceforge.net/lists/listinfo/grinder-use">https://lists.sourceforge.net/lists/listinfo/grinder-use</a><br>
</blockquote></div><br>
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
grinder-use mailing list
grinder-use@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/grinder-use
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic