an Internet-scale operating system does not yet exist--developers have already produced a number of Internet-scale, or peer-to-peer, applications that attempt to tap the vast array of underutilized machines available through the Internet [see box]. These applications accomplish goals that would be difficult, unaffordable or impossible to attain using dedicated computers. Further, todays systems are just the beginning: we can easily conceive of archival services that could be relied on for hundreds of years and intelligent search engines for tomorrows Semantic Web [see "The Semantic Web," by Tim Berners-Lee, James Hendler and Ora Lassila; Scientific American, May 2001]. Unfortunately, the creation of Internet-scale applications remains an imposing challenge. Developers must build each new application from the ground up, with much effort spent on technical matters, such as maintaining a database of users, that have little to do with the application itself. If Internet-scale applications are to become mainstream, these infrastructure issues must be dealt with once and for all.We can gain inspiration for eliminating this duplicate effort from operating systems such as Unix and Microsoft Windows. An operating system provides a virtual computing environment in which programs operate as if they were in sole possession of the computer. It shields programmers from the painful details of memory and disk allocation, communication protocols, scheduling of myriad processes, and interfaces to devices for data input and output. An operating system greatly simplifies the development of new computer programs. Similarly, an Internet-scale operating system would simplify the development of new distributed applications.
http://www.sciam.com/2002/0302issue/0302anderson.html
