A comparison of several performance test tools is made here, including: Jemeter,pylot and Mul-mechanize. Not a deep user, using these three tools/frames in a day, write down a little bit of view here.
A. Brief introduction:
- Jemeter:java platform under the old-style performance testing tools, almost around the HTTP protocol as the core of a tool, full-featured;
- Pylot: A python-based platform for testing Web server performance and extensibility tools
- Mul-mechanize: A python-based platform for testing Web server performance and extensibility tools
Two. Features
- Jemeter:
1. Simple process as long as familiar with the HTTP process, understand their own use case scenario, can quickly get started;
2.Sample can be connected to each other, so it is a bit like a function, each function is a sample, functional decomposition of the good, jemeter can achieve some of the more grand features. Unexpectedly, Jemeter can actually do the text information crawler. (If the result tree can be extended to other formats, it can do more things);
3. Distributed testing, this has not been done, but when the client performance is not enough, this feature is very useful;
4. Recording function, this is useless, do not evaluate;
5. Can be integrated in the Jenkins, to achieve the performance of the automated construction
6. Cross-platform, both Windows and Linux can be used, but Linux is not very useful
- Pylot:
1. This thing can be large, flexible expansion, if you need to get drawing results, installation Wxpython
2. The test scenario is set by Testcases.xml , and then the parameters are configured by the command line. It can be said that Testcases.xml determines the quality of use of Pylot, there is a certain threshold
3. There are python places where there is pylot, so cross-platform is also sure
- Mul-mechanize:
1. If you compare Jemeter to Django, then mul-mechanize must be flask.
2. Extensible, if you need to draw the results, you can install the Matplotlib library yourself
3. The threshold is relatively high because scenarios such as get,post requests and assertions need to be implemented in the Python language in the transaction class, but this also shows its flexibility.
Maybe that's all, after studying in depth, I'll summarize.
Comparison of several performance test tools/Frameworks
