How to do click-through rate test in website stress test? The more detailed the better, thank you! !

See how many visits, response speed, fault tolerance rate and other performance indicators can be carried by the website web server shortly after its completion. These are what managers want to know and are most worried about. How can we know all this? It is a good method to conduct Web stress testing through tools. It can effectively test the running status and response time of the Web server.

I. Test environment:

Hardware: CPU: Athlon XP 2500+, memory 5 12MB, hard disk 80GB.

Server operating system: Windows Server 2003

IIS: 6.0

BBS: Dynamic Network 7.0

IP: 192. 168 1.20

Tools: Web application stress tools

Second, introduce tools.

There are many tools available for Web stress testing, such as Microsoft's Web application stress, siege under Linux and full-featured Web-CT, all of which are excellent tools for Web stress testing. Although these tools are convenient for us to test the endurance of the server, they are "double-edged swords". Attackers can use any comprehensive testing tool to launch a catastrophic denial attack on a small Web server.

Next, the author takes Microsoft's Web Application Stress Tool (WAST) as an example to conduct a Web stress test. This is a set of tools specially developed by Microsoft website testers for actual website stress testing. Through this powerful stress testing tool, managers can test the website in a real environment before the website actually goes online, so as to find out the potential problems of the system and further adjust and set up the system.

Third, tool settings.

Download and install WAST, the process is extremely simple. Then run WAST, and you can see that its interface is very simple. Before stress testing the target Web server, you should first make some necessary settings for it.

1. Sets the number of parallel connections.

Click Default → Settings on the left to open the Settings panel. Set up parallel connection under concurrent connection. The thread with the smallest stress level (thread) and the thread with the largest stress multiplier. Threads here specify how many threads the program uses to make requests in the background, that is, how many client connections are simulated. Generally fill in 500 ~ 1000, because the number of threads is set according to the tolerance of this machine. If you have enough confidence in your machine configuration, you can set it higher. (Figure 1)

Figure 1

2. Set the duration

It is used in "Test Run Time" to specify the duration of stress test, which is divided into days, hours, minutes and seconds. For example, we set it to 1 hour. (Figure 2)

Figure 2

3. Other settings

"Rpquest Delay" sets the delay time, which we set to 100~500. The "Pause" setting sets the pause time, the warm-up time is the initial test running time, and the cooling time is the test time at the specified end stage. Bandwidth "specifies the bandwidth bottleneck, which allows you to simulate the network bandwidth from a modem of 14.4 Kbps to a local area network (LAN) connection of 1( 1.5 Mbps). Redirect sets the redirection time, Throughput sets whether to enable user and password page state saving, and Name Resolution sets whether to perform name resolution. The above options can be set according to your own needs. (Figure 3)

Figure 3

Fourth, the pressure test

After the setting is completed, the pressure test can be carried out, and the test steps are as follows:

Step 1: Click the "new script" button on the toolbar and click the "Southern" button in the open panel to create a new test project. Set it in the window that opens, and fill in the IP address of the server to be tested in the main options. Here we fill in192.168.1.20, and choose the following test Web connection mode. Here Verb chooses get, Path chooses the path of the web page to be tested, and here we fill in /Index.asp, which is the home page file of the dynamic network. Trash can (Figure 4)

Figure 4

Step 2: Set the pressure level (thread number) to 1000 in the function setting of "Settings". When finished, click the gray triangle button in the tool to test. During the test, we can see from the task manager of the server that the CPU utilization rate has reached 100% and the loss rate has reached the maximum, as shown in Figure 5. Using the command netstat -an in the CMD window, you can see that the IP address of the client has been connected to port 80 on the server. As shown in Figure 6, the website cannot be opened, prompting too many users to connect. (Figure 5) (Figure 6)

Figure 5

Figure 6

Summary: Through the Web stress test, the administrator has a general understanding of the compression resistance of the Web server, so that the server hardware can be expanded according to the actual needs, and it also provides a basis for system setup and software selection. To sum up, it is very necessary to carry out stress testing before the Web server is officially released.