Apache and IIS have not been returned. Xitami has arrived.

Article title: Apache and IIS have not returned Xitami. Linux is a technology channel of the IT lab in China. Includes basic categories such as desktop applications, Linux system management, kernel research, embedded systems, and open source.
Currently, open source code is dominant in Internet Web server software platforms, cross-platform Apache HTTP and Microsoft Windows NT/2000 IIS
　　
. It seems that they will continue to dominate the market in the next few years. However, other competitors are eager to try, including cross-platform servers.
　　
One of these is the new Xitami Web server, which is said to be a small, fast, and powerful multi-platform open source Web server. This product will be released in two ways: the free version and the Xitami Pro version. The difference between the two versions is that the Pro version only adds an SSL layer, so it supports SSL/2, SSL/3, and strong TLS encryption.
　　
In this article, we will discuss the installation on Windows and Linux platforms. I will only demonstrate the free version of this software. Some standards will also be introduced to compare this server with the current industry leader.
　　
Install on Windows
　　
I installed Xitami on a single-processor Windows 2000 host. To achieve the purpose of the experiment, I chose the Turbo installation method to achieve the highest execution efficiency. In fact, in the Microsoft platform environment, the installation is very simple: Download the executable file, double-click the icon after the download is complete, and then just answer a few questions.
　　
After the installation is complete, you need to create two simple files: defaults. cfg and defaults. aut. The ults. cfg file determines where the system can find the file containing the user name and password. defaults. aut contains the actual file name and password. Defaults. cfg is generally like this:
[Security]
Filename = defaults. aut
　　
The ults. aut file is as follows:
[Security]
Filename = defaults. aut
　　
As you can see, defaults. cfg points to the file named defaults. aut to verify the user name and password. These two files are stored in the C: Xitami directory. Because the server will re-read this directory in 60 seconds, you do not need to restart the system. This configuration is mainly completed by a Web-based management tool.
　　
After that, Xitami can start to work. You will notice that there is a small, green ring containing X in your system. You can start, stop, and configure Xitami.
　　
Standard
　　
All three servers are built on Pogo, and the Athlon 750 contains mb ram.
　　
. I start, detect, and close a server at a time. I use the httperf standard software to generate a total of ten tasks (each task includes five calls and each call interval is two seconds) at the speed of one task per second in each individual test ).
　　
If you want to perform a more rigorous test, you can use this command to make httperf generate 1000 tasks at the speed of one task per second (each task includes 100 calls, every call interval is one second ).
　　
Figure A is the result of the first Test.
　　
　　
　　
Image
The results of the first Test were quite obvious. You can see that the results obtained on the Apache host are similar to those on Windows and Linux. IIS is also outstanding, similar to the results of the Linux Apache server. Xitami is another conclusion: the Xitami server won the competition.
　　
For example, observe the request rate data. On Windows Apache, each request takes 21.4 milliseconds and 46.6 requests in one second. In Linux Apache, the speed is significantly improved. each request only takes 11.0 milliseconds and 91.1 requests per second.
　　
Skip these steps and directly view the Xitami. the speed increase is exponential. On Windows, Xitami only takes 5.5 milliseconds for each request, so that 180.7 requests can be processed in one second.
　　
Finally, Linux Xitami requires 5.5 ms and 181.1 requests in one second. From the slowest (Apache running in Windows) to the fastest (Xitami running in Linux), the speed has increased by 74.3%. But what is the execution scale? To test this problem, I decided to send a large number of requests to these servers from many other machines at the same time.
　　
To implement this experiment, a large number of commands are sent from two machines to each server and the results are recorded. For ease of viewing, I only keep the results on one machine (the same machine is used from start to end), which is listed in Figure B.
　　
　　
Graph B
　　
If you do not consider the platform, this test is successful on these servers. We note that this result is not as balanced as it was for the first time. Although the hardware and platform environment have not changed, the results have been completely changed due to increasing the number of commands and joining another client. IIS is the only server with a small change in results. Although it cannot catch up with the Xitami, IIS performance is relatively stable.
　　
Does this result show that Xitami has nothing special about it? In this scale test, we see a decrease in efficiency. To verify this conclusion, I use another command mode on both IIS and Xitami servers and the result is in figure C.
　　
　　
Graph C
　　
As you can see in the results of the third Test, the Xitami server continues to lead the IIS server in most aspects. Even if you observe the CPU time, the two servers will display the same result. This is because Xitami is converted from Turbo to Normal. Before conversion, the CPU usage of Xitami is 100%, and that of IIS is only 50%. Why? After carefully checking the Xitami configuration, I found that the difference lies in the efficiency of configuration execution. IIS and Apache run as backend servers. When the Turbo environment is used, because the Xitami server is always running on the front end, all CPU cycles are consumed as quickly as common applications.
　　
Despite high execution efficiency, it is not the best for your hardware. If you are willing to discard expensive servers, you will be willing to visit the Xitami installation page. you can click the Web-Based Admin link on the Xitami homepage of your server or go to the website http://server.ip.addy/admin.
　　
On the Web-Based Admin page, click Configuration and click Advanced in the new window. On the Advanced Configuration page, you will see the Server Performance configuration entry. Select Background, click Save, and then click Restart. Now your server will run as a backend server, which will reduce execution efficiency, but will save a lot of CPU clock cycles.
　　
Summary
　　
Xitami is worth noting. After testing, it performs very well. Its speed, size, simple management method, and price advantages will make it a powerful competitor in the Web server market, which can well meet the needs of your Web site. (Translation: Zhi De; responsible Editor: Liu Yanzhi)
　　
TechRepublic is an online community and resource center designed for various IT professionals. we provide in-depth technical articles written by IT experts to address IT expert readers. In addition to Windows, e-mail, firewall, and other articles, we also provide IT industry analysis, downloads, management points, discussion groups, and news and email services.