search

w3f / messaging

Messaging for Web3
169 stars 12 forks source link

Decentralised public key infrastructure (PKI) #22

Open gpestana opened 5 years ago

gpestana commented 5 years ago

PKI infrastructure seems to be one of the bottlenecks in terms of scalability and decentralisation for secure and anonymous P2P networks. Is there any research work done at the moment to tackle this?

In the context of this messaging protocol, I can think of relying on external (decentralised) PKI network, although I'm not even aware that such thing exists nowadays.

burdges commented 5 years ago

There is necessarily some portion of the PKI that resembles the accounts database in a blockchain like Ethereum, so the problem is:

  1. users should be lite-ish clients,
  2. who sample uniformly, but
  3. do not reveal anything abut their sampling,
  4. and the whole database rolls over say once per moth.

Initially, you can ignore 1 to make all users semi-full-nodes, like almost all existing serious anonymity system do, including Tor. If you instead ignore 2 and 3 then you cannot claim to bemore secure than Tor in all respects.

At a later stage, addressing all concerns requires using an anonymity layer in bulk sampling. If you use the mixnet itself then you have a circularity problem. If you use PIR then you have the problem of making PIR decentralized.