1. 16
  1. 7

    Have a peer-to-peer web, but keep the concept of servers in addition to it. It’s no better a solution than currently exists, but at least it offers potential redundancy (fueled by popularity, but perhaps some other metrics can be devised, as was hinted towards the end of the article) instead of the current lack of existence that information has a tendency to vanish into.

    Additionally, I’d like to point out that things such as paywalls or pizza delivery would still require servers that are not data-addressable, because they are services and not information that can be copied and distributed (at least, not without a clever reworking of how the service works). Servers are here to stay for the time being, and a peer-to-peer web most likely won’t change that dramatically.

    1. 5

      BitTorrent has web seeds (download from HTTP servers in addition to peers). They’re quite popular: used by many OS distrubutions, Archive.org, media.ccc.de, Amazon S3

      1. 1

        By “keep the concept of servers”, do you mean that whoever publishes to the network should also self-host their data and treat the redundancy as a bonus? This way of looking at it has also occurred to me. You could describe this as a more robust form of client-side caching, where all static assets are saved locally and the validity of the cache is communicated through the network protocol. Thinking of the p2p web as a bandwidth optimization makes sense to me in exactly the way you say “it’s no better a solution than currently exists, but…” we get some nice additional properties at scale. But this point of view falls short of the ambitions the decentralized web movement has for these network architectures.

      2. 1

        Pay people to store things.

        A crypto currency which mines not blockchain but content would encourage people to donate their disks to the rare- hoovering up all the worlds data, since anyone who wants it would have to pay a premium inverse to availability. Think of it as a tax-on-demand library of Congress.

        1. 3

          This wouldn’t work by itself. Soon, most disks would be occupied by useless junk and someone will need to decide what to ditch and what to keep. Which is the other function of the Library of Congress.

          Some library evaluation methods include the checklists method, circulation and interlibrary loan statistics, citation analysis, network usage analysis, vendor-supplied statistics and faculty opinion.

          – Wikipedia on Collection development

          1. 1

            Soon, most disks would be occupied by useless junk and someone will need to decide what to ditch and what to keep.

            I’m envisioning a recurring storage fee that would eventually run out unless topped-up. Somewhat like Ethereum distributed apps that stop running when they run out of ‘gas’.

          2. 2

            FileCoin aims to be that. Its initial token sale raised over $200M, showing that a lot of big players want in on that market opportunity. Right now it seems they are massively expanding their team, and it’s not clear yet when it will be available to the general public.

            Considering P2P rewards, private torrent trackers have been doing this for a really long time, converting seed time into virtual community credits or something similar, enabling recognition and opportunities to contributing members. But like in other parts of the online world, spending time, money, and equipment for a cause rather than a product becomes less and less convenient for the average user. Many people lamented the downfall of what.cd, but it illustrates the two sides of the P2P coin pretty well: it can have huge potential if many people are willing to invest their resources, but it is still very much illegal for much of the shared content, and there is a powerful force behind the corporations and authorities to stop these things (namely, huge piles of money).

            1. 1

              A reward system like you describe appeals to me. A market can be an efficient way to allocate a finite supply of resources. This would also enable things like bounties for data that exists out of band. I wonder if valuing the data inversely proportional to its availability would eventually bring about an equilibrium where most things were within the same range of availability. I also agree with the sibling that storage space is a complicating factor. In theory, the value of the data would rise and attract more hosts until the supply met the demand. So the effect would be a general pay-wall.