In progressPerformance TestingSometimes, the response time tested using the performance tool LR is longer than the time that IE is actually used. For example, when you use IE to open a system, you only need 1 ~ 2 seconds, and the result of running a user using LR may be 8 seconds, 10 seconds, or a larger response time.
There are three situations to analyze and summarize the above problems:
1. Think time and log recording time are not ignored during LR running;
2. ** or the server's machine configuration is not high. For example, if the system resource of a Low-configuration machine is full when the tool is running, the response time is too long.
3. The actual ie experience time is not the same as the LR recording response time.
In the first two cases, you can use LR settings and improve hardware configuration.
In the third case, the response time of a transaction during LR recording includes DNS resolution, connection, first buffer time, receive time, and client time, for example, when we use IE to open the page, the system will first perform domain name resolution, establish a connection with the server, and download data. At this time, the page can be displayed in IE, however, the actual response time is not over, and the browser may still interact with the server, or the client IE does not send the request in time due to busy times, client latency occurs (client IE will execute some JavaScript scripts or other page initialization actions), until these actions are completed, it is a complete response time, LR is also the response time recorded.
Therefore, we usually use IE to experience less time than the response time recorded with performance tools. Therefore, the response time on the system page should be based on the time recorded by the tool, and the average response time should be viewed in the analysis report.
Toone
--------
Explanation of time:
1. DNS resolution: When browsing and accessing a website, the domain name is generally used. The DNS server needs to resolve the domain name to an IP address. This process is the domain name resolution time.
2. Connection: After the IP address of the webserver is parsed, the browser request is sent to a Web server, and an initial HTTP connection must be established between the browser and the web server, the server needs to do two things: receiving requests and allocating processes. The connection is established.
3. First Buffer: After a connection is established, the first data packet is sent from the Web server and transmitted to the client over the network. The time when the browser successfully receives the first byte is first buffer.
4. Receive: receives the first byte from the browser until the last byte is successfully received and the download is complete.
5. Client time: the latency of requests in the client browser.
When LR is used for page request response time, I used to download the page and clear the IE cache. Each time different users browse the page, the pressure on the server is relatively high, the number of requests per second is high, but the response time is indeed much larger than the manual time. I personally think that during the loading process of images and CSS style files, they are all loaded synchronously, however, for the IE client, the person's vision will first see the image, and then the appearance of the style, so it feels faster, because it is a dynamic process, but for LR, it only cares about the time from the first request to the end of the request, so the time will be longer than the actual operation time.
For example, it may be an image link, but it may be a hyperlink to be loaded, but this link is implemented using Ajax and CSS is also included, but for IE end users, as long as you see the image, the loading is complete!
Note: The LR response time is inconsistent with the time used by IE.