TorrentFreak reports on WebTorrent, a project using BitTorrent and WebRTC to transport files:
WebTorrent is a project launched by Feross Aboukhadijeh, a Stanford University graduate who has already booked quite a few successes in his career. After graduating he founded PeerCDN, a P2P-assisted content delivery network, which was sold to Yahoo at the end of 2013. Feross then focused on WebTorrent, convinced that it could revolutionize how the web works today.
"I felt that the idea of 'people-powered websites' – websites that are hosted by the visitors who use them – was too revolutionary to keep locked up as proprietary software, and I wanted to do more to push the idea forward," he tells TF. "Imagine a video site like YouTube, where visitors help to host the site's content. The more people that use a WebTorrent-powered website, the faster and more resilient it becomes."
[...] "WebTorrent is the first torrent client built for the web. It's written completely in JavaScript – the language of the web – and uses WebRTC for true peer-to-peer transport. No browser plugin, extension, or installation is required," Feross tells TF.
Over the past two years WebTorrent has matured into a project that's slowly starting to win over several major tech companies. Netflix, for example, contacted Feross to discuss his technology which they may use to stream their videos. A few months ago Netflix specifically mentioned WebTorrent in a job application, which shows that the video giant is serious about P2P-assisted delivery.
[More After the Break]
Feross believes that companies such as Netflix could benefit greatly from WebTorrent. Currently, streaming performance goes down during peak hours but with WebTorrent this shouldn't be a problem.
[...] Netflix aside, there are already various noteworthy implementations of WebTorrent. The project's homepage, for example, shows how easily it can stream video and βTorrent offers a fully functioning torrent client UI.
Other examples include File.pizza, which uses WebTorrent to share files in the browser. The same technology is used for server-less websites by PeerCloud and Webtorrentapp, while GitTorrent uses it to decentralize source control.
In addition to the examples above, the Internet Archive is also looking into the technology for its video distribution, and another major tech company is considering adding WebTorrent support to their web browser.
(Score: 2, Troll) by FatPhil on Monday December 14 2015, @10:39AM
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 2) by takyon on Monday December 14 2015, @01:51PM
Is it really abuse to prioritize beginning of files (like uTorrent can already do with torrent streaming)? You can give higher/lower priority to different files in a torrent, why not different pieces of a file as well?
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 1) by loic on Monday December 14 2015, @03:22PM
+1
How is selecting preferably some blocks more than others is abusing the system? About every single torrent client allows to prioritize files, or even to skip some specific files, and it could probably be extended to specific blocks. The bittorrent protocol does not specify any download order AFAIK. The only prioritizing rule I know of is that it tries to distribute first the less available blocks, yeah, an upload rule.
Oh my god, they use an impure proprietary fork of bittorrent! Burn them!
(Score: 2) by FatPhil on Monday December 14 2015, @04:07PM
I think worst-case behaviour means that after O(N) transfers only O(sqrt(N)) of the N pieces have left the seeder if there are sufficiently many (Omega(sqrt(N))) badly-behaved leechers. Random behaviour would raise this to O(N). Not just that, but as all similarly-bandwidthed peers will get the same blocks at about the time, there will be way more 'have' messages that are uninteresting, so you're decreasing the available bandwidth.
Sure, a few clients having a preference for a particular subset of a torrent, such a track from an album, can be done, the algorithm is resilient to that. However, for the scenarios presented *every* client will be having *exactly* the same preferences, and all of the mathematical modelling that prove the original works efficiently have to be thrown in the bin.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 0) by Anonymous Coward on Monday December 14 2015, @04:50PM
In a streaming media scenario, the fast downloader will probably stay connected until the video finishes playing (unless the stop halfway through).
I can envision such downloaders seeding for as much as 90% of the time.
If many people stop watching the video half-way though (due to the content, not technical reasons), then the first portion of the video is most important anyway. Presumably trying to "seek" through a boring video will also distribute preferred blocks to some extent.