Communal Hosting: Using the torrent protocol to host websites
The Torrent Web
I have been wondering for a while; would it be possible to host a website using the torrent/magnet protocol? Surely, the torrent protocol probably isn't adequate as it is right now a modifications to it could allow it host a website.
Let's say that we used the magnet link as a base (which is still a torrent after all). You create an alternate version which we would call a "HyperMagnet". The only different with the Magnet link and the HyperMagnet link would be that the magnet link automatically redirects the user to a torrent client while the HyperMagnet, knowing that the file it is downloading is a webpage, would stay within the browser. This would allow users to "browse" torrent files.
Of course, the website wouldn't be able to simply depend on the community; a main server would still be needed. This is why a new protocol, based on the torrent protocol, would be needed. What would need to happen for things to work would be a a system included in the browser that allows the mothership to update the "drones" every day at midnight for example or even hourly if the files aren't too big.. The mothership would seed the update to 500 main drones, which would would then seed to 2,500 drones, with would then 50,000 drones, etc... Some type of hierarchy would be needed so that the users who are willing to seed a lot are chosen as the main drones, secondary drones, until the chain ends.
This would mean that some of the changes would not be updated instantaneously such as real HTTP servers but for many websites such as Demonoid or TPB, hourly updates would suffice.
Seeding vs. leaching
For this type of system to work, seeding would have to be mandatory to use to server. Considering the number of users using websites such as TPB, we can't have only 10,000 people seeding while 2M are leaching. If you were to use the system, you would have no choice but to seed as well.
Now, I don't know much about handling servers so this is just an idea but it seems that with a little bit more thinking, someone could make it work in the near future. Downtimes would be a thing of the past, server costs would decrease by thousands for many websites (using the bandwidth most regular users never use), and much more. This, alongside with an alternative DNS, would be ever many people's dream come true.