| || |
Open source Performance tool for Desktop Application Cherwell 8.0.2
Can anyone suggest good OPEN SOURCE PERFORMANCE tool for DESKTOP APPLICATION to test minimum of 100 users and maximum of 200 users
There are no actionable requirements related to the architecture of your application, the environment in which the tool has to run, the analysis and reporting requirements of the tool or the skills or your performance testing team to be able to make a recommendation on any tool, open source or commercial.
Thanks jpulley3. Cherwell is user interface for both Desktop and Browser Clients. Its Incident Management, Problem Management, Change Management tool... Can anyone recommend any open Source Performance tool
And what is the communications architecture between the client and the server which you need to reproduce?
And what are the monitoring requirements for your infrastructure?
And what are the analysis & Reporting requirements for your test?
And what platform does this test tool need to run on?
And what are the skills of your testing team including programming, systems analysis, etc...? The mechanical bits of the testing tool are only 5-15% of the skills and knowledge you need to effectively pull off a high value test?
If you want someone to go pull the architecture for Cherwell and make a recommendation then be prepared to fork over a couple of K USD for consulting to help you figure out what tools will even work
Notice what is not on the evaluation criteria? License model, Open Source or Commercial. The reason? If it doesn't meet your business, technical or skill requirements then it doesn't matter what tool you pick, the value will still be negative.
On the other hand, if you have solid programming and performance testing skills in house, then it does not matter which open source tool you pick, for you can modify the source to meet your test production, reporting and analysis requirements. Be sure to track the labor required for this modification and compare it to the cost of the commercial tool however. You may be surprised what you find when compared to a short term rental option.
First of all, I would suggest contacting Cherwell support or customer service, there is a non-zero chance they have already done the performance testing so you don't have to waste your time and efforts, i.e. looking into here it seems to support 500+ concurrent users (however the test scenario sucks in my opinion)
If you still looking for a tool, here is a list of the most advanced free and open source load testing tools currently available:
- Apache JMeter
You can review the comparison and figure out whether any of them is applicable and if yes which one is better.
Agreed that the test definition is very lacking as it appears to minimize number of users for the preference of a specific type of incident being created. Especially when compared against the standard for a benchmark definition, the TPC-C Benchmark. http://www.tpc.org/tpc_documents_cur...-c_v5.11.0.pdf
Originally Posted by email@example.com
Thanks everyone for your inputs... I'll check on your inputs.
Myself coming from not knowing your stack or technologies used, it'll help frame the problem in the following ways..
1. What are you testing?
- Are you testing the client, server, database, or middleware (or combinations)? Ex: client receiving streaming updates from a server such as movie streaming vs. server receiving data from many clients such as typical rest request.
2. What protocols are you testing over?
- Different tools have support for different protocols. Many tools works with http based protocols out of the box (REST, SOAP/XML), but some protocols like raw TCP and raw UDP is less supported.
3. What is the level of logic you need in your tests?
- Dumb single path tests can be done using record and playback tools like many of the commercial tools out there or JUnit. Tests requiring conditional and looping logic you may opt for a more of a parallel programming approach such as locust.io or mechanize.
I know if no performance test which produces actual value which uses the model above, independent of the license model of the tool. You can tell this code when developed as such. When I see it I recommend immediate termination for the resource.
Originally Posted by dlai
Avoid Record and Playback as a model.
Are you trying to start a flame war by just quoting a single phrase out of context and throwing some quippy comment? Fact is, a majority of load test involve some level or recording and playback with parameterization. The core of what any load/performance testing tool does is throw data at an application. And the most common way of doing that is replaying a request/response sequence with a parametrization. I doubt that you can recommend everyone who does that gets terminated without looking silly.
Originally Posted by jpulley3