5 ms·

To me, this signals a strong need for more secure bitcoin storage options, and since this is hacker news, perhaps some of us should get started on that :)

drivebyacct2 14y ago
Paper wallets are easy to make and relatively easy to use. Slightly harder to spend money with. I suppose someone could customize a small Linux live CD to boot, not touch the hard drives, load an encrypted wallet off a secure jump drive and encourage backups. It's fairly straightforward if you use a dm-crypted jumpdrive and, say for example, Tails.
danielweber 14y ago
I think I could do a pretty good job keeping wallets secure. But it would need to be a full-time job, not a side project, and it would need a big budget. You don't bootstrap a bank. (Plus I did some basics with Bitcoin a few years ago and got pissed off at the protocol and moved onto less annoying things.)
- M4v3R 14y ago
  I don't like to spam with links, but http://www.bitalo.com http://www.bitalo.com aims to be the service you described, and will be launched soon. "Most secure" really means that no one, even the site admins/hosting platform can never touch your coins. This will be enforced by the technology used, not just some internal policies. And also, the site will be backed by a german AG company, which is basically a type of "Public limited company" backed by minimum of 50,000 EUR.
  - foobarqux 14y ago
    How do they plan to accomplish this without using javascript crypto?
    - M4v3R 14y ago
      I plan on using Javascript crypto. I've spent many hours reasoning over this idea, also studying security community reactions on mega.co.nz and I think it is possible to do now. I don't want to disclose all details now for obvious reasons, but all I can tell that I will not reinvent the wheel here - all parts needed are already available and mature, you just have to assemble them into a complete solution.
      - foobarqux 14y ago
        The point is that you are overstating the security benefit of your solution. A breach of the server will now only compromise wallets of anyone who logs in until the breach is detected instead of all wallets instantly.
      - epsylon 14y ago
        > you just have to assemble them into a complete solution. Many (if not most) of the security vulnerabilites of the past years come from perfectly safe components assembled in an unsafe way. Crypto-engineering is hard.
  - drivebyacct2 14y ago
    No one can never touch my coins? Uh oh! Blockchain.info can't touch my coins. But if there servers are compromised, a hacker could inject a tiny, tiny amount of JS and have my ID/password sent to... anywhere... and then the hacker could access my account. I'm curious to know how you'll get around that vulnerability.
    - M4v3R 14y ago
      The security on the client side will be on the level that blockchain.info provides, but this will be more than just a online wallet - it will also be an exchange. I'm aware about injecting JS vulnerability. Of course you can't get around it with anything on the main server. It's possible however to setup an external server that will be monitoring the files and firing alarms the minute something's wrong (asset checksums doesn't match). And that's exactly what we'll do. Another thing is that all SSH/SCP access is also logged and the whole team gets an email immediately when it happens.
stcredzero 14y ago
I may be spending a little karma in saying so, but this is exactly what DRM would be most useful for, when utilized by individuals in their own interest, as opposed to corporations using it to exercise power over individuals. This is exactly where being able to detect if an iPhone is non-jailbroken would be useful. Or, we can dispense with having DRM on hardware we personally won. How about a protocol for providing a secure bitcoin wallet in the cloud? Basically, everyone runs their own open source mini-OS in their own obfuscated VM (1), which a smartphone app sets up for you and to which a smartphone app only acts as an interface. The wallet information is never on the smartphone, and the DRM is never on your own hardware. Such an infrastructure would have many other uses, not just bitcoin. (1) The security of each VM instance would be on an economic basis. Each one might be breakable within a week, say, but the system could be set up in a such a way that the week old information isn't work enough to motivate the effort.

drivebyacct2 14y ago

Paper wallets are easy to make and relatively easy to use. Slightly harder to spend money with. I suppose someone could customize a small Linux live CD to boot, not touch the hard drives, load an encrypted wallet off a secure jump drive and encourage backups. It's fairly straightforward if you use a dm-crypted jumpdrive and, say for example, Tails.

danielweber 14y ago

I think I could do a pretty good job keeping wallets secure. But it would need to be a full-time job, not a side project, and it would need a big budget. You don't bootstrap a bank. (Plus I did some basics with Bitcoin a few years ago and got pissed off at the protocol and moved onto less annoying things.)

M4v3R 14y ago
I don't like to spam with links, but http://www.bitalo.com http://www.bitalo.com aims to be the service you described, and will be launched soon. "Most secure" really means that no one, even the site admins/hosting platform can never touch your coins. This will be enforced by the technology used, not just some internal policies. And also, the site will be backed by a german AG company, which is basically a type of "Public limited company" backed by minimum of 50,000 EUR.
- foobarqux 14y ago
  How do they plan to accomplish this without using javascript crypto?
  - M4v3R 14y ago
    I plan on using Javascript crypto. I've spent many hours reasoning over this idea, also studying security community reactions on mega.co.nz and I think it is possible to do now. I don't want to disclose all details now for obvious reasons, but all I can tell that I will not reinvent the wheel here - all parts needed are already available and mature, you just have to assemble them into a complete solution.
    - foobarqux 14y ago
      The point is that you are overstating the security benefit of your solution. A breach of the server will now only compromise wallets of anyone who logs in until the breach is detected instead of all wallets instantly.
    - epsylon 14y ago
      > you just have to assemble them into a complete solution. Many (if not most) of the security vulnerabilites of the past years come from perfectly safe components assembled in an unsafe way. Crypto-engineering is hard.
- drivebyacct2 14y ago
  No one can never touch my coins? Uh oh! Blockchain.info can't touch my coins. But if there servers are compromised, a hacker could inject a tiny, tiny amount of JS and have my ID/password sent to... anywhere... and then the hacker could access my account. I'm curious to know how you'll get around that vulnerability.
  - M4v3R 14y ago
    The security on the client side will be on the level that blockchain.info provides, but this will be more than just a online wallet - it will also be an exchange. I'm aware about injecting JS vulnerability. Of course you can't get around it with anything on the main server. It's possible however to setup an external server that will be monitoring the files and firing alarms the minute something's wrong (asset checksums doesn't match). And that's exactly what we'll do. Another thing is that all SSH/SCP access is also logged and the whole team gets an email immediately when it happens.

stcredzero 14y ago

I may be spending a little karma in saying so, but this is exactly what DRM would be most useful for, when utilized by individuals in their own interest, as opposed to corporations using it to exercise power over individuals. This is exactly where being able to detect if an iPhone is non-jailbroken would be useful. Or, we can dispense with having DRM on hardware we personally won. How about a protocol for providing a secure bitcoin wallet in the cloud? Basically, everyone runs their own open source mini-OS in their own obfuscated VM (1), which a smartphone app sets up for you and to which a smartphone app only acts as an interface. The wallet information is never on the smartphone, and the DRM is never on your own hardware. Such an infrastructure would have many other uses, not just bitcoin. (1) The security of each VM instance would be on an economic basis. Each one might be breakable within a week, say, but the system could be set up in a such a way that the week old information isn't work enough to motivate the effort.