In computing, and specifically , seeding is the uploading of already downloaded content for others to download from. A peer, a computer that is connected to the network, becomes a seed when having acquired the entire set of data, it begins to offer its upload bandwidth to other peers attempting to download the file. This data consists of small parts so that seeds can effectively share their content with other peers, handing out the missing pieces. A peer deliberately chooses to become a seed by leaving the upload task active once the content has downloaded. The motivation to seed is mainly to keep the file being shared in circulation (as there is no central hub to continue uploading in the absence of peer seeders) and a desire to not act as a parasite. The opposite of a seed is a leech, a peer that downloads more than they upload. Seeding is a practice within peer-to-peer file sharing, a content distribution model that connects computers with the use of a peer-to-peer (P2P) software program in order to share desired content. An example of such a peer-to-peer software program is BitTorrent. Peer-to-peer file sharing is different from the client–server model, where content is directly distributed from its server to a client. To make peer-to-peer file sharing function effectively, content is divided into parts of 256 kilobytes (KB). This segmented downloading makes the parts that peers are missing be transferred by seeds. It also makes downloads go faster, as content can be exchanged between peers. All peers (including seeds) sharing the same content are called a swarm. Data shared via peer-to-peer file sharing contains shared file content, computing cycles and disk storage, among other resources. In peer-to-peer file sharing, the strength of a swarm depends on user behaviour, as peers ideally upload more than they download. This is done by seeding, and there are different motivations to do this. There are two popular motivations to seed, of which one is the reputation-based incentive mechanism and the other is the tit for tat mechanism.
Michael Christoph Gastpar, Su Li
Rachid Guerraoui, Anne-Marie Kermarrec, Maxime Monod, Davide Frey