On 14 May 2012 18:48, James A. Donald <jamesd@echeque.com> wrote:

Some design characteristics:

*. Peers are identified by a durable public key.

*. Peers interact with keys they do not trust, but expect protocol conformance and prosocial behavior only from keys they trust. Peers limit their exposure to non trusted keys so that they can suffer only limited damage from non trusted keys deviating from protocol or otherwise misbehaving.

*. Peers may not only trust Bob, but transitively everyone Bob trusts, and everyone that those peers trust ... If this works, they increase their transitive trust in Bob, if it fails they decrease their transitive trust.

*. Prosocial behavior consists of conforming to protocol, storage availability, with valuable, or at least demanded, stuff in storage and accessible on demand, and paying ones bandwidth and storage debts. Since networks are unreliable, perfect protocol conformity is never expected or demanded, but the burden that one's deviations from protocol place on other peers counts against one's karma.

Hm. I'm not sure how to bootstrap when you only want to interact with people you trust - but it seems like this could be a model for a Web of Trust that grows over time if you have a way to validate what people are supposed to be sending. I am Alice and I interact with 1000 other people, Bobs1-1000, who send me pieces of a file. They're valid, so I sign their public keys saying they were good for some number of bytes, send them the signature, but also keep it locally. I send data to 50 other people, Carols1-50 who in turn give me signatures. Over time, I interact with Bob50 again, and trust him based on my previous signature. I find a Dave, who I trust a little because he has a signature from Bob600, and Bob600 was good to me. The key pieces (and why this couldn't work for distributed DNS) are that I know that Bob[x] sent me is valid because I have the information to build a merkle tree and validate his piece. The more I download, the more people I 'trust' because they were honest actors last time. The more I upload, the more people 'trust' me because I acted honestly in the past. But I may have reinvented something simple or missed an obvious flaw, because I'm not well-versed in how DHT works, or anything published in this area. -tom _______________________________________________ p2p-hackers mailing list p2p-hackers@lists.zooko.com http://lists.zooko.com/mailman/listinfo/p2p-hackers ----- End forwarded message ----- -- Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org ______________________________________________________________ ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE