Use Application Center Test (ACT) for stress testing

After we have completed sps2003-based development and implemented our specific applications, can we Directly invite users to use them? If I do this, experienced developers will not even do stress testing! I really don't want to live ......

 

Oh, yes. In the development environment, only functions are taken into account. In a specific environment, when a large number of users access these functions, will many functions fail? What about performance ?...... Let's take a look at how to perform stress testing.

 

I believe that there are certainly many tools for stress testing, but most of them are currently used in Application Center Test (ACT ). This is a component of vs. net. It is very simple and easy to use. It also supports scripts and implements complex functions ......

 

The test procedure is omitted here. Assume that we test a website directly, for example, test. sendrequest ("http: // server/default. aspx "). How can we analyze the results?

 

The following are some of the information I have just learned and shared with you. I hope that you can use an interesting feature for experienced users.

 

1. First, check whether there are any errors. For example, 401 users do not have any verification errors. If there is an error, the result is definitely incorrect and you don't need to read it.

 

2. Analyze average requests per second, which should be "average requests per second ".

 

We can test it several times, using 1, 2, 5, 10, 50, 100, 200 ...... Number of concurrent browser connections. Then, we can analyze the results in a chart several times.

 

Generally, as the number of concurrent browser connections increases, the number of average requests per second increases, the increase causes the average requests per second to decrease. Then, this value is almost the maximum number of concurrent browser connections supported by the server.

 

3. Average time to last byte

 

The time when the server response is received after the request is sent.

 

Obviously, this value increases as the number of concurrent browser connections increases. Generally, the following criteria can be used to analyze whether the value is reasonable:

 

0.1 seconds: very fast

1 second: the speed is still very fast, so you don't need to consider performance issues.

3-4 seconds: an acceptable result for the internal network

5-8 seconds: an acceptable result for external networks

Over 10 seconds: too slow

 

4. Average time to first byte

 

In the same way, it is only the time when the server starts to respond after the request is sent.

 

You are welcome to discuss and share your experience with Act.

From: http://blogs.msdn.com/oliverlu/archive/2004/12/21/328273.aspx