Topic: VCFS: A novel approach to dealing with disadvantages of cryptocurrencies (Read 951 times)

if it deals with money you should never trust anybody.
in the vm image there could be an pre installed app which escapes from the vm and uses host network. afaik such attacks are possible with most vms.

If the computer does not connect to network at all, any OS would be fine

as someone who has paid to have a filesystem built for a bulk-storage service (cyphertite online backup), i can tell you that having a FS just for cryptocurrencies is a bad idea. creating and maintaining an FS is a ton of work without a whole lot of gain in this case.

existing FSes work just fine when storing and manipulating a blockchain. the blockchain is already its own sort of userland filesystem.

its not hard to put a linux distro together and distribute it as a vm image.
but i would never use one myself: it just requires to much trust in you (as the distributor) and your hosting (could be hacked and image is replaced with another one).

Thanks I did.

I think the best way to go, if the forum thinks this is a good idea worth pursuing, is to use the Linux from scratch distro and make a small virtual machine that boots in single user mode and runs headless with only a network interface for communication. The goal would be a tiny virtual box that can run on any host machine Windows/OSX/Linux and act like a distributed form of Trezor.

Not a trivial task by a long shot but I'd love to take a stab at it. But I would like to see how much interest there is for something like this.

Have a look at Othercoin or Trezor.

Hello,

  I have an idea I would to present, depending upon the feedback I will devote more time to it.


I propose that many of the basic problems with virtual currency stems  from the fact that they are stored on a general purpose file system. All file systems are built around the concept that beyond some basic I/O permissions a file is a block of data on a storage medium which can be easily copied, edited and moved. There is an embedded idea of trust in the end user behind this. While this is a highly desired design goal for 99.999% of all computer applications I would propose that it is not desirable for a crypto-currency.

I believe that the basic ethos of an ideal crypto-currency is nobody can be trusted, only machines can be trusted. The current file system is designed to provide open information and trusts the end user. In an ideal crypto-currency system when I transfer money from my bank account to create a virtual coin it is owned by system, while I can control how and what it is spent it on it is the property of the system until I spent it on a tangible good. I can’t be trusted beyond that, since I could be a malicious user. I can’t be trusted to know where it is, but only the guaranteed that it does exist and I can perform financial transactions with it.

I propose a file system that violates many of the current rules of software abstraction. A general purpose file system manages data on a hard drive as a utilitarian application that is a servant to other applications. This file system would be an embedded system that contains an application, networking and disk management all in one.
It could almost qualify as a mini-virtual machine.

This file system would offer a unique interface to the user, not a directory tree but a small menu of options to be manipulated by the bit coin client. The file system would be able to communicate with other file systems using virtual synchrony to safely transfer data, distribute computing and provide a backup in case of a hardware failure. The location of data would be hidden to the end user, the ability to copy a wallet at the user level would not exist. If a wallet was lost the user can request from the system a recovery wallet from stored fragments on the network, this process would of course be safe guarded to prevent a false claim. 

By creating a massive hidden distributed file system the need for intense encryption is rendered partially mute. The unorthodox and complex nature of such a system would be its own form of encryption on top of any traditional RSA encryption.

Any thoughts? (Other then I'm crazy )