P2p

From Hackepedia
Revision as of 02:42, 16 November 2006 by Pbug (talk | contribs)
Jump to navigationJump to search

A Peer-to-Peer network as we know it is often used to distribute content on the Internet where the sender lacks money but the content has enough value for people to give a helping hand in distribution, or the sender moves around a lot meaning he cannot access a central site easily. Peer to Peer networks are set-up over unicast where multicasting is not available.

How it works is that you "seed" data and the first person to download this becomes the second person to help distribute the same data. It cascades (think atomic chain reaction) into a larger group that offer content and the last person to download this data is going to enjoy the best download speeds. Contrast this to broadcasting and multicasting which has infrastructure in place to allow only one person to send to a group of people (one to all vs. all to one).

This author wants you to consider "buscasting" or "daisycasting". Basically it's a P2P network but the nodes are arranged in a daisy-chain meaning each node has 2 connections one being a reciever and the other a transmitter. At the end of the chain is a transmitting node which serves an audio stream. When time of an audio broadcast isn't of essence when it reaches a listener (such as elevator music, or just something to drown the silence, which causes some people to feel uncomfortable) buscasting can be done on any ISDN or DSL broadband connection with appropriate compression. It is up to the chain to decide whether the broadcaster unicasts checksums to every peer individually in order to ensure the audio content is the same at the end of the chain as at the front, the penalty is that as the chain grows so does the traffic to the source.