Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.

Mirroring is the process that exactly duplicates information in one location to another. It can be done locally-say, between disk drives in the same system-or globally, such as when information on a server is duplicated to a server in other parts of the world. Many Web sites mirror their content to other servers, bringing information closer to users, and reducing the distance and number of router hops that data must travel to get to a user. Content distribution is all about getting Web-based information closer to users.

Most network operating systems and many desktop operating systems support disk mirroring. In this type of mirroring, stored data on a primary drive is continuously copied to a second storage device in real time so that both devices hold the same information. Mirroring is a form of fault tolerance that protects data from equipment failure. There are several different types of mirroring, as described here and pictured in Figure M-2.

• Mirroring    Data is copied from on-disk controller (channel) to two disk drivers. If one drive fails, the other is still operational.


• Duplexing    Data is duplicated over two disk channels and stored on two drives. This method extends fault tolerance to the controller.


• Server duplexing    This method provides fault tolerance by duplicating the entire file server. If one server fails, the other provides continuous service to users. For example, Novell's System Fault Tolerance provides server duplexing.


• Replication    A strategy of duplicating critical files and directories from a server at one location to a server at another location to make that information more accessible to users at the remote location and also to provide redundancy and backup. See "Redundancy" and "Replication."


• Clustering    A cluster is a group of servers that share access to the same resources and service clients equally. Should one of the servers go down, the others take up the processing load. Clustered servers may access the same disk systems, which may be mirrored or in a RAID configuration. See "Clustering."


• Mirror site    A mirror site is a duplicate data center, located at another site, that contains duplicate systems and data. The duplicate data center should go into operation as the primary site if the master data center site fails for any reason. Companies running mission-critical applications will often create mirrored sites. See "Data Center Design."

Many organizations today outsource in order to obtain the benefits and features of the mirror techniques described earlier. They may also choose to co-locate duplicate equipment at public Internet data centers. Outsourcing and/or co-location allows organizations to gain all the management and security benefits of Internet data centers. See "Outsourcing."

Content distribution is an advanced, highly managed form of mirroring. It improves performance of Web sites by placing copies of content at caching servers close to users. These servers are placed at ISP sites and Internet data centers by companies such as Akamai. One can imagine a ring of content distribution devices surrounding the Internet. When users access Web sites, the content they are looking for is likely to be cached in a content distribution server near them. See "Content Distribution." Also see "Web Caching."

Copyright (c) 2001 Tom Sheldon and Big Sur Multimedia.
All rights reserved under Pan American and International copyright conventions.