Load Testing tool
I am using WebLoad, but I find that it is not as good as Load Runner.My company is not ready to buy LoadRunner.
Has anyone used MS WAS tool which is load testing tool?
If anyone has used ,please let me know how this tool is.
Re: Load Testing tool
Here is a summary sent by Mr. Falafel to comp.software.testing. I agree with him that MS WAS is not the perfect load testing tool, but it is better than not having any tool at all.
It's free and our programmers find it very useful for what they are trying to accomplish, during the unit and integration testing phases.
During our system testing phase we will be using Performance Studio (since it's already in house).
>Subject: Re: Web App Stress Tool Query
>From: Mr Falafel email@example.com
>Date: 3/13/00 7:29 AM Pacific Standard Time
>In article <firstname.lastname@example.org>,
> email@example.com wrote:
>> I'm very new to Web Site testing and have been charged with the task
>> figuring out whether or not the MS WAS is up to the job. I'm having
>> problems though because the site I'm trying to test is secure and
>> are authenticated via certificates.
>> Does anyone have any experience of using the WAS in these conditions?
>MS WAST is a benchmarking tool that is trying to be pawned off as a
>proper stress tool. It's freeware that has no support and has a number
>of issues with it. We dismissed it as a proper test tool very quickly.
>It's great for benchmarking server/hardware configurations and that's
>about it. If you want to stress your web app as actual users would,
>you're better off getting one of the real tools (webload, Loadrunner,
>Here's some findings from various sources (some from this group)
>Here's my findings. I can get more dirt on this product if need be but
>I think this makes it clear:
>WAST appears to be a scaled back load test tool capable of generating
>basic load tests against standard web based, three tiered applications.
>It is really a type of WCAT tool that allows scripting & session
>recording. WAST is most commonly used as a sophisticated benchmark tool
>in the development and proving of web-architecture configurations and
>tunings, especially useful for ASP pages.
>In the consideration of using it as a full service load test tool,
>there are a couple of issues that should be pointed out:
>The product is 'freeware' and while user-feedback is
>encouraged, 'Microsoft does not guarantee a response or product
>The tutorial does not recommend more than 1,000 virtual users, or
>scripts with more than 1000 lines. This limits the scalability of the
>VBA(VB) style scripting language has very few features that
>PerformanceStudio has. No parameterization or randomization. Even
>simple features like wait times have to be manually entered.
>Maintenance of these scripts would be very labor intensive. The level
>of expertise needed by the test team in VB, ASP, server configuration
>is very high. The learning curve of this tool would be very steep and
>knowledge transfer would be difficult.
>The tool cannot run more than one script at a time, except on
>separately configured machines. This does not reflect actual user
>patterns and just stresses the web server. The reporting of this type
>of test would very inconclusive.
>Web Application Stress is a raw protocol level request generator, where
>as PerformanceStudio is an application level load generator.
>Web Application Stress plays back tests scripts sequentially at the
>protocol level; hence it is unable to reproduce realistic use of the
>application by real browsers that can send several requests in parallel
>for each virtual user.
>Web Application Stress requires you to understand and specify your
>tests in terms of the underlying protocol knowledge of HTTP is required
>in order to build the tests.
>Pages that require dynamic data require that you manually edit the
>query string name-value pairs and change the POST data.
>You cannot test applications that require session information to be
>carried forward from one page to another. Only cookies are supported,
>but not any of the other ways to handle a session Id (such as including
>it as part of the URL, or storing it in a form).
>Pages coming back are not automatically validated (only the first byte
>and last byte of data are accounted for). If the server returns a page
>that has in the HTML, Server Busy, it is treated as a valid response
>and the load test continues without any errors.
>Web Application Stress has very rudimentary scheduling and no
>capabilities to dynamically increase the load while the test runs.
>You cannot graph on the performance data only generate plain text
Senior Author "Automated Software Testing"