Editing GiFT (section)

== OpenFT protocol ==
[[Image:Apollon screenshot.png|thumb|The Apollon front end]]
giFT's sibling project is OpenFT, a [[peer-to-peer]] file-sharing [[network protocol]] that has a structure in which nodes are divided into 'search' nodes and 'index' [[supernode (networking)|supernode]]s in addition to common [[node (computer science)|nodes]]. Since both projects are related very closely, when one says 'OpenFT', one can mean either one of two different things: the OpenFT protocol, or the implementation in the form of a plugin for giFT.

Although the name ''OpenFT'' stands for "Open FastTrack", the OpenFT protocol is an entirely new protocol design: only a few ideas in the OpenFT protocol are drawn from what little was known about the [[FastTrack]] protocol at the time OpenFT was designed.{{fact|date=February 2018}}

=== OpenFT file-sharing protocol ===
Like FastTrack and [[Napster]], OpenFT is a network where [[node (networking)|node]]s submit lists of shared files to other nodes to keep track of which files are available on the network. This reduces the bandwidth consumed from search requests at the price of additional memory and processing power on the nodes that store that information. The transmission of shared lists is not fully recursive: a node will only transmit its list of shared files to a single search node randomly chosen as that node's "parent", and the list of those files will not be further transmitted to other nodes.<ref>From the giFT documentation at [http://gift.sf.net the giFT homepage]</ref>

OpenFT is also similar to the [[gnutella]] network in that search requests are recursively forwarded in between the nodes that keep track of the shared files.

There are three different kinds of nodes on the OpenFT network:

; USER
: Most nodes are USER nodes; these have no special function.
; SEARCH
: These nodes handle search requests; they search the filelists their CHILD nodes (explained below) submit to them. These nodes must have a capable Internet connection and at least 128M [[Random Access Memory|RAM]]. A modern processor is highly recommended as well.<ref>According to the giFT documentation:<blockquote>Search nodes handle search requests. They search the filelists their CHILD (common) nodes submitted to them. These nodes must have a capable Internet connection and at least 128M RAM. A modern processor is highly recommended as well.</blockquote></ref>
; INDEX
: Nodes with fast connections and much memory can be INDEX nodes, which keep lists of available search nodes, collect statistics, and try to maintain the structure of the network.<ref>According to the giFT documentation:<blockquote>INDEX nodes keep lists of available search nodes, collect statistics, and try to maintain the structure of the network.</blockquote></ref>

A node can be both a SEARCH and an INDEX node.
USER nodes will pick three SEARCH nodes to be their PARENT nodes. They will submit their shares list to them if the PARENT accepts the USER as its CHILD. By default, SEARCH nodes will be PARENTS for a maximum of 500 CHILD nodes.