Napster was introduced in 1999 and is historically the first P2P system to achieve global-scale deployment, reaching
50 millions users in just one year. It comprised a central server which stored the index of all files shared by the
participating peers. To locate a file a user queried the central server using the name of the file and received as a result
the IP address of a peer containing the file. A direct connection was established between the requesting peer and
the peer containing the file in order for the download to be effected. In other words, although file downloads were
performed in a P2P fashion, resource location was centralized. The central index server used in Napster is not easy to
scale and provided a single point of failure. Although Napster is historically considered as the first unstructured P2P
system to be deployed, the existence of a central index differentiates it considerably from today’s unstructured P2P
systems. Napster has been characterized as “the fastest growing Internet application ever”.
In today’s unstructured P2P systems, each peer maintains a constant number of connections to other peers, called its neighbors, thus an overlay network of peers is formed. Gnutella and KaZaA are considered two of the most popular unstructured systems. Due to the lack of an underline structure in those systems, there is no information about the location of files, thus the prevailing resource location method is a broadcast-like process called ”flooding”.
Top of the page
Print this page | Add this page to your favorites
Copyright © 2005 VTT