Wiki-torrents

This is a proposal for a new Wikimedia sister project.

Creating wiki-torrents
Status of the proposal
Status	rejected
Reason	no interest in many years. Pecopteris (talk) 06:41, 19 August 2023 (UTC)

The storage and dissemination of data in the form of torrents can solve several problems at once:

• Decentralized storage of data that will provide invulnerability Wikipedia in case of any problems with the central servers.
• Offload server.
• In some cases, much faster access to some resources, especially large.

Unlike Wikipedia resources from other types of distributed resources of the torrent is that the objects of Wikipedia (articles or parts of articles, attachments) may be changed to the original source (central server). Therefore, wiki torrent client should receive notification of a change of objects handed out, and in the event of changes exclude the object from the distribution until it is updated from a central server. Data integrity of the torrent system is checked by computing the hash stored files. In the setting of a wiki torrent client, you can specify limits on the amount of data storage and upload speed, CPU load. In accordance with user central server caches the user's computer a piece of data.
Probably reduce the load on the central server will significantly reduce the cost of maintenance for the Wikipedia community. In addition, in the event of data loss central servers (up to destroy the central server) and the problems of recovery from backups Wikipedia will continue to operate successfully. Caches of data stored in user, you can quickly re-create the temporary (and in the case of destruction of central servers even permanent) central server.
Weak Link offers: development of methods for handling web-browser to the torrent networks. Likely to have to develop a browser extension that allows for accessing the resource Wikipedia (page file) to request the resource (and its components) in the first place in the torrent network, rather than on a central server. But it is preferable to do without extensions, for example using a client-oriented tools java: java on, there are several projects working with torrents (for example, Java Bittorrent API). Or, the ability to download the requested resource object can provide installed on your computer wiki torrent client. Needless browser extension can be used as a Wiki-BitTorrent client.
The system of two complementary duplexing servers (central and torrent trekkers) will be more reliable than the use of one central. You can use the system multitrekkera: one torrent tracker is tightly integrated to the central server to Wikipedia, other / Gia (backup) is located separately. Module on the client (in the form of the same java-code) in case of unavailability of the torrent trekkers primary server can switch to the backup. Tight integration to the central server Wikipedia provides query performance of the torrent network to a resource (object) to a central server that is comparable to a direct request object http. In this case, object faster starts with a central server, but if the object is large enough and loading is time-consuming, or a central server is overloaded, then the time to connect to other seeds, and leaves a central server.
The torrent includes storing information on the content of knowledge Wikipedia and not user pages. Torrent module should only help load the content pages (mainly media content) and interface elements. This preserves not the whole page, and in the form of individual dumps blocks: Interface + content (text and attachments). A torrent system will provide stability and speed "surfing" on Wikipedia in view mode. Authorization same (with editing capabilities) will be possible only if the central server of Wikipedia.

Proposed by

Lebedev Igor V

Mailing list links

http://stranger-live.livejournal.com/10053.html

People interested

1. Jesus Presley (talk)
2. Sovereign Sentinel (talk)