(no title)
MorphisCreator | 10 years ago
Random of those blocks, based on your nodes ID and the hash of each block's data, are stored on your computer. Before they are, your node encrypts them and throws away the key. All that is ever stored on your harddrive is random encrypted data and no key. The key is the original key that the uploader is provided with. The encryption key is the hash of the data, the ID of the data is a hash of that key. Your node knows only the ID, it cannot derrive the key to decrypt the data. Your node cannot be tricked into storing unencrypted data.
Now, a soon future version will support seeding v1. This is the ability to essentially 'like' a file/website/etc, and that will ensure that your node never throws out the data blocks it does store for that file/site/files_that_site_links_directly. This is mostly UI work that is missing. This feature will also automatically seed for example your Dmails, comments, trust publications, own site, Etc. (Although a UI will let you control it).
After that, more distant, will be seeding v2. This will be more like the file sharing that is similar in concept to how https://peeriodproject.github.io/dl/peeriod_an-anonymous-app... envisions file sharing. Their paper came out as I was coding and I noticed they were the closest in many ideas to mine of any papers I've seen. Their project seems stalled, as most coming out of the academic world are. That is why I started with code and not a paper :) I will write the paper before 0.9 release.
MorphisCreator|10 years ago
That relation of your node's ID is one of the Sybil proofing measures designed early on.
I designed everything to be transparent and modular with crypto so I can switch to better algos as they become available (quantum proof, I am watching you!). The first addition will likely be ECC once I have decided upon a safe (as in non NSA backdoored) ECC curve which I believe the community has not proven yet as safe as RSA-4096.