Had this thought the other day and tbh it’s horrifying to think about the implications of one, or God forbid all, of them going down.
Stackoverflow too but that only applies to nerds haha

  • csm10495@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    3
    ·
    1 day ago

    In theory this could be true. In practice, data would be ripe for poisoning. It’s like the idea of turning every router into a last mile CDN with a 20TB hard drive.

    Then you have to think about security and not letting the data change from what was originally given. Idk. I’m sure something is possible, but without a real ‘omph’ nothing big happens.

      • csm10495@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        1
        ·
        5 hours ago

        Hashed by whom? Who has the source of truth for the hashes? How would you prevent it from being poisoned? … or are you saying a non-distributed (centralized) hash store?

        If centralized: you have a similar problem to IA today. If not centralized: How would you prevent poisoning? If enough distributed nodes say different things, the truth can be lost.

        • Possibly linux@lemmy.zip
          link
          fedilink
          English
          arrow-up
          1
          ·
          5 hours ago

          This is a topic that is pretty well tested. Basically the data is validated when received.

          For instance in IPFS data is tracked by its hash. You request something by a CID which is just a hash.

          There are other distributed networks and they all have there own ways of protecting against attacks. Usually an attack requires a huge amount of resources.

          • csm10495@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            1
            ·
            3 hours ago

            Even in ipfs, I don’t understand discoverability. Sort of sounds like it still needs a centralized list of metadata to content I’d, etc.