Decentralized Storage Comparison
jdb10987 at yahoo.com
Tue Jan 29 13:39:33 PST 2019
Interesting, useful development. Looking at the table, in the second item below, I see that most of these storage systems use "client side" encryption. I assume that means that the person who enters the data is in charge of the encryption. But I can see the possibility that this information would be not intended to be encrypted, yet the people (thousands? millions?) of them who might eventually participate in storing data would not want to experience negative consequences from storing what, to them, is a mass of what are (to them) unknown documents, potentially of questionable legal origin.
Suppose one of the (millions?) of users of this system (ones who want their documents stored) decides to not encrypt some challengable document. (and by "document", I don't mean just text; pictures or video would apply; for concreteness, let's say certain forms of illegalized pornography, or perhaps copies of the famous DNC emails publicized just before the 2016 election.)
Would that document appear, in the clear, on the computer of dozens or thousands of people who allow their computers to be used for this Decentralized Storage? Remember, commonly-available hard disks exceed 10 terabytes these days. It is inconceivable that the actual owner of the storage node could reasonably ensure that this system could be "clean" of potential legal and jurisdictional problems. Yet, it is quite conceivable that the powers-that-be might want to harass participants in this Decentralized system? They could do so simply by storing a lot of in-the-clear challengable documents, monitoring their locations, and then legally harassing people whose computer contain that material.
In principle, governments may not actually object to the presence of that data, Rather, they might simply use this as a nexus to harass people they otherwise object to. In other words, such harassment could be very selective.
One solution (of course) involves encryption, but not necessarily the way you'd expect. Think back to the Vernam Cipher (aka "one time pad") https://en.wikipedia.org/wiki/Gilbert_Vernam What if a challengable document, call it "A", is essentially split up into two: Take random (or pseudorandom) string, the length of document "A", call it "B", is XOR'd (exclusive-or'd) with "A", and the result we will call "C", of the same length as "A" and "B". Then, instead of having document "A" stored, store both "B" and "C", but maybe not on the same storage nodes. Basically, an implementation of a one-time pad. Or, instead of merely two strings, this could be expanded, in principle, to any number.
The purpose of this is not to conceal the ultimate information, but to split up that information so that no one operator of a storage node contains enough information that arguably violates the law in the jurisdiction he happens to be at. WIll this work? Laws can be changed, but it would be difficult for a law to prohibit someone from possessing data that could conceivably be combined with some other information, somewhere, in order to regenerate some banned document "A".
On Tuesday, January 29, 2019, 11:40:27 AM PST, grarpamp <grarpamp at gmail.com> wrote:
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 4725 bytes
Desc: not available
More information about the cypherpunks