Each subsequent interaction is driven by what the site owners decides to write in the JavaScript code that was rendered in the browser.When users access your site, they download the torrent pointed by your public key via the DHT, and immediately query it using sqltorrent by asking for the site->index field, and render it in a browser.If you’re writing a search-engine, you should use SQLite’s Full Text Search tables. The database may also contain other data - which can be queried from your JavaScript page.The database should contain a table site, and a field index, which contains the blob of the HTML you’re sharing - you can bundle all of the CSS/HTML/JS together, or have them as separate fields in the table.Your public key points to a torrent which shares a single SQLite database file.33cwte8iwWn7uhtj9MKCs4q5Ax7B) that are shared and kept alive using the BitTorrent network via the Mutable Torrents extension ( BEP46) which uses the DHT network. Torrent sites are accessible via your public key (eg.People download pieces in “rarest first” order so the swarm (all the peers downloading a specific torrent) remains healthy.Ī possible TorrentNet site that crawls the DHT and TorrentNet sitesĪlthough most of the ideas are there, gluing all these pieces together into a single “browser” with a simple user experience is complicated. The idea of splitting files into pieces allows for a better distribution of the file across a large number of peers.

Rather you “ask for pieces” from different peers that may have that data. In BitTorrent when you download a file, you don’t download it sequentially like you would download a file from, say, an HTTP server. The main idea is that a site can be packaged as an SQLite database, shared as a torrent, and can be queried on demand without having to wait for the entire torrent to be downloaded. I set on a quest to solve this issue and that’s why I created TorrentNet. P2P (peer-to-peer) sites have gained much attention lately with systems such as ZeroNet and IPFS, which seem to improve on older systems like Freenet.īuilding search-type-functionality - where users type words in a text box, and results appear - on top of these distributed systems is not quite feasible yet because users need to download the entire site (usually several hundreds of gigabytes large) before they can run queries against it.

Bit torrent search