"Secrets" of the existence of nerdy
Wikipedia is one of the top 10 sites in the world. It has about 500 million unique visitors per month . This is about 200 requests every second! Wikipedia is a fully non-profit structure, donations are its main method of financing. This means that MediaWiki , which is the basis of Wikipedia, not only keeps the site in the top, but also does it with very modest means. And therefore, work on performance, caching and optimization is put at the forefront.
“Server idle time is our favorite way to make money,” jokes Wikipedia’s productivity engineer, Domas Mituzas, “the point is that while servers are not available, the donation request for the new equipment is displayed on the page. Being the most sought-after non-profit structure, Wikipedia is a good example of high performance web site. While Google and Microsoft can spend half a billion dollars to build a data center, Wikipedia is quietly located on a modest number of servers in the data centers in Tampa and Ashburn. There are also servers in San Francisco and Amsterdam.
')
“The traditional approach to site accessibility is not exactly our choice,” Mituzas said at the O'Reilly Velocity conference , “and I’m not saying that you should do the same. But losing a few seconds is not critical for our cause. And as long as we do not face serious difficulties, there will be no problems. ”
And let Wikipedia engineers are not very serious people, they take their work very seriously. In accordance with the principles of Wikipedia, which recognizes the priority of society over commerce (on the site we will not meet advertising!), The resource receives a good income from donations. Wikipedia supports almost 100% availability. And these are not all staggering numbers :
Also, who are interested, interactive statistics are available, where you can find data on different versions of Wikipedia, including Russian.
The site appeared in 2001 and worked on the CGI Perl script, hosted on the same server. By 2008, Wikipedia already had about 200 application servers, 20 data servers and 70 Squid cache servers. At the moment, the number of servers has grown to 885. It works on MediaWiki software, which is written specifically for it, and now it is an open source project. MediaWiki uses PHP based on MySQL. In addition to Squid, Wikipedia uses Memcached and a load balancer for Linux virtual servers ( LVS ). Wikipedia also uses shard applications to establish relationships between databases. More technical details can be found in the presentation of Mituzas and Mark Bergsma from WikiMedia.
Mituzas summed up his vision of Wikipedia’s work in a blog , namely in a post about the Velocity presentation: “I think in this context Wikipedia is interesting as an example of not quite successful work - a tiny budget, a bold approach to infrastructure, a conservative development method, deception and a bunch of cheap tricks (caching, caching and caching again) ".
As long as there will be such a resource on the Internet as Wikipedia, we can confidently say for whom the first position of search results will be assigned to no less erudite Google and Yandex.
PS A brief history excursion
“Server idle time is our favorite way to make money,” jokes Wikipedia’s productivity engineer, Domas Mituzas, “the point is that while servers are not available, the donation request for the new equipment is displayed on the page. Being the most sought-after non-profit structure, Wikipedia is a good example of high performance web site. While Google and Microsoft can spend half a billion dollars to build a data center, Wikipedia is quietly located on a modest number of servers in the data centers in Tampa and Ashburn. There are also servers in San Francisco and Amsterdam.
')
“The traditional approach to site accessibility is not exactly our choice,” Mituzas said at the O'Reilly Velocity conference , “and I’m not saying that you should do the same. But losing a few seconds is not critical for our cause. And as long as we do not face serious difficulties, there will be no problems. ”
And let Wikipedia engineers are not very serious people, they take their work very seriously. In accordance with the principles of Wikipedia, which recognizes the priority of society over commerce (on the site we will not meet advertising!), The resource receives a good income from donations. Wikipedia supports almost 100% availability. And these are not all staggering numbers :
- 23.130 million registered users;
- 34.246 million pages (English version);
- 8.790 million views per hour (English version);
- 4,701 million article (English version).
Also, who are interested, interactive statistics are available, where you can find data on different versions of Wikipedia, including Russian.
The site appeared in 2001 and worked on the CGI Perl script, hosted on the same server. By 2008, Wikipedia already had about 200 application servers, 20 data servers and 70 Squid cache servers. At the moment, the number of servers has grown to 885. It works on MediaWiki software, which is written specifically for it, and now it is an open source project. MediaWiki uses PHP based on MySQL. In addition to Squid, Wikipedia uses Memcached and a load balancer for Linux virtual servers ( LVS ). Wikipedia also uses shard applications to establish relationships between databases. More technical details can be found in the presentation of Mituzas and Mark Bergsma from WikiMedia.
Mituzas summed up his vision of Wikipedia’s work in a blog , namely in a post about the Velocity presentation: “I think in this context Wikipedia is interesting as an example of not quite successful work - a tiny budget, a bold approach to infrastructure, a conservative development method, deception and a bunch of cheap tricks (caching, caching and caching again) ".
As long as there will be such a resource on the Internet as Wikipedia, we can confidently say for whom the first position of search results will be assigned to no less erudite Google and Yandex.
PS A brief history excursion
Source: https://habr.com/ru/post/243159/
All Articles