[iancoleman]

The technical side of this project is pretty interesting.

There's TOR-style anonymity using Reliable Message Delivery [0] (and to a lesser degree also Secure Message Delivery [1]). The whole routing aspect of this network is quite the fascinating rabbit hole.

Sybil attacks are addressed with Node Age [2] rather than proof of work, which should be very energy efficient but is maybe open to bribery or collusion (still an open question).

Sharding (to use the ethereum term) is implemented using Disjoint Sections [3] which splits the content-addressable data into sections so nodes only store part of the total data, making it much more efficient from a storage point of view than blockchains like bitcoin or ethereum1.

Unlike most cloud storage (eg storj, s3 et al) the network does not use erasure coding, instead relying only on redundancy. It'll be interesting to see how this plays out since there's a lot of existing research detailing the benefits of erasure coding [4].

The economics of the network are still very far from being finalised and will be challenging to get right. One of the key motives for the project is free-to-access (so anyone can browse) so that means storage nodes are not directly rewarded by clients for the necessary bandwidth to serve that data. It's not clear exactly how this is going to work but given the success of torrents (which are not directly rewarded for seeding) it seems that this particular motive should be possible to achieve.

This is just the technical aspects I think HN people will find interesting. The social impact of truly permanent data and truly private user-owned data will also be unique aspects of this network.

[0] https://github.com/maidsafe/rfcs/blob/bebf53550bb2171deadb76...

[1] https://github.com/maidsafe/rfcs/blob/bebf53550bb2171deadb76...

[2] https://github.com/maidsafe/rfcs/blob/bebf53550bb2171deadb76...

[3] https://github.com/maidsafe/rfcs/blob/bebf53550bb2171deadb76...

[4] https://maisonbisson.com/post/object-storage-prior-art-and-l...