Central servers of large projects often do not have the data transfer rate to make their services available to all interested parties in a reasonable time. For this reason, the data is "mirrored" on other servers, so a 1: 1 image is created and made available to interested parties there.
Since the synchronization usually takes place asynchronously , this “mirror image” is not always up-to-date, but it is an important method of keeping data available on the Internet at all times. For example, the Domain Name System is essentially based on “mirroring” the data known to the servers. The update does not take place in real time , but at specified time intervals .
For example, there are copies of all Linux distributions on several servers on the Internet. In this way, the data is available for download on different computers, and the resources are distributed when accessing heavily frequented data. The respective mirror servers are, for example, synchronized with the current data from the source every night at 3 a.m. (mostly with rsync ).
Linus Torvalds describes an urban legend as the author of the following saying as a comment on the fact that the loss of his source texts for the Linux kernel would be intercepted by mirroring them from several computers:
“Backups are for wimps. Real men upload their stuff on the internet and let the world mirror it. "
“Backup copies are for cowards. Real men upload their stuff on the internet and let the world mirror it. "
Example with Wget
With the free command line program Wget , Internet pages can be mirrored with the option --mirror or -m.
wget --mirror --convert-links --backup-converted --html-extension -o log http://www.gnu.org/