Web application stress tool (was, web application load testing tool)

Your Web How many concurrent users can access the server? You have ever encountered a server DDoS Is the attack paralyzed? Here we will introduce the famous website stress testing software developed by Microsoft website testers, Microsoft Web application stress tool (was, web application load testing tool ), And it's free. Its: http://download.microsoft.com/download/a/8/2/a82e7ba7-c772-4ec4-b186-2cf147f42c11/setup.exe Was Is a website Performance Test and Evaluation Software. It simulates a large number of concurrent users to access the server at the same time to obtain the server's affordability. Software like this is a double-edged sword, depending on which aspect you use. If it is useless, it will cause some losses to your server. If it is used, you can find out how much load your server can withstand in a timely manner. In order to take appropriate measures in a timely manner. To test the website load, you must first create Was Script to simulate user access and other activities. How to Create a script: Record browser activities; Import IIS Logs; By Was Point Web Website content; Or you can make it manually. Here I use part of the script generated by recording browser events, I. Preparations before testing 1. Clear temporary files such as cache and cookies of other websites of IE browser before testing. Ii. Test Script Creation 1. Open was and click record2. check the activity to be recorded. 3. Click finish4. a new browser window will pop up automatically to record your browsing content. Then, access the Web page you want to test. 5. When you access your server, was records these activities. After the access is completed, click stop recording. 6. On the script page, you can view the collected scripts and enter the IP address of the server in the server column. 7. delete the World of Warcraft gold8element with a low latency. You can press Ctrl to select multiple elements at the same time, and then click Delete in the toolbar to delete 9. click Settings. Here, you can set parameters such as the number of initiated connections, warm-up time, bandwidth limit, and how long the test will take. Stress Level And Stress Multiplier These two items determine the number of concurrent connections to the server. Level indicates the number of threads produced by the client. A thread can generate multiple concurrent socket requests. Therefore, the value obtained by multiplying the two values is the number of concurrent connections initiated by the client at the same time. . Microsoft recommends this Stress Level The value cannot exceed 100 . If the number of concurrent connections to be simulated exceeds 100 , Can be adjusted Stress Multiplier Or use multiple clients for testing. The time settings include "test run time", "request delay", and "Suspend. The test run time is set in the format of day, hour, minute, and second. It is recommended that the test run time be too short. If the set concurrency is large, the time should increase by comparison, in order to generate enough requests, and the stop time means timeout when the connection exceeds this time; Warmup WOW Gold (Warm-up) time, generally can be set 1 ~ 3 Minutes. In Warmup Period Was Start to execute the script, but do not collect statistics. Warmup Time MTS And database and disk buffering. If Warmup Collects statistics within the time period. The overhead of these operations will affect the performance test results. Bandwidth limit ( Throttle bandwidth ). The bandwidth limit function can be simulated for testing. Modem ( 14. K , 28.8 K , 56 K ), ISDN ( 64 K , 128 K ) And T1 ( 1.54 m . The bandwidth limit function can accurately predict the customer's access through a dial-up network or other external connections. Web Server performance. III. test pressure 1. click Start in the toolbar. 2. click Run test to start testing. 3. testing server pressure. The time left is the time remaining 4. if you are prompted that the RPC service is unavailable, check whether the PRC service is enabled and set it to automatic. in this case, you can use netstat-An to check whether many connections have been enabled. The established in the State column indicates that connections have been connected. 6. check that most of the CPU processes are occupied by WebTool 7. view CPU usage 8. use the performance monitor to view the current performance. view server performance 4. view test report and analyze data 1. click Report on the toolbar to view the test report. 2. Focus on "socket errors" and "result codes. Socket errors is divided into connect, send, Recv, and timeouts. Connect indicates the number of times the client cannot connect to the server; send indicates the number of times the client cannot correctly send data to the server; Recv indicates the number of times the client cannot correctly connect from the server; timeouts indicates the number of timeout threads. If these four values are small, or even 0, the server can withstand the test. If the value remains high, or even close to the set concurrency, check your server. " Result codes. If the value in the code list is 200, all requests are successfully returned by the server. If the value is 400 or greater than 400, for example, 404, in this case, you need to find the "page data" node on the left side, view the specific error project, and modify the project accordingly. 3. speed 4 connections, Server IP address,. Time Parameter 5. it is recommended that multiple clients perform the test. Do not set a large number of connections on one test machine. stress level value , which are distributed in different places. When you test the server's Intranet, the webpage cannot be displayed, and the webpage accessing other websites cannot be opened. In this case, you can try your server if you are not a friend in the same LAN. Continuously increase or decrease the number of connections. After multiple tests, you can see how much pressure the server can withstand. If the server is set up in IIS, you can also modify the maximum number of connections allowed. After obtaining the data, analyze the data, server resource distribution, response processing speed, what measures should be taken when a large number of users or are under attack, and performance optimization.