Bitcoin Forum

Hey all, i remember reading about different types of tx's that the BTC protocols supports more that just simple P2P tx;s.  That is, i know there is an escrow tx.  But were BTC to take off and large businesses were to use it, how it the world could they maintain security over their bitcoin balance?  That is to say, what's really stopping some kid in the accounting dept from txing all the available coins to an address he controls and then fleeing the country?

I imagine that control software can be devised that would hold the keys in a special sort of way, but is there anything built into the protocol?  That is, can we have an account that would require more than one signature to reduce the risk of this?  For example, an address that requires authentication of 3 out of 5 people to sign a tx.  Then coins could be diversified amongst accounts that require different circles of folks consent to send the coins.  Also it would be best if it didn't require ALL people because people leave, die, become nasty and hold out etc..

Corporation right now largely rely on a traceable/reversible banking system.  The worry more about detecting misappropriations than they have to worry about recovery techniques.  Once a higher up in a corporations treasury tells the merchant bank that a transfer was illegal they reverse it quickly.  No way to do that with BTC.    You have to worry about prevention instead of just detection.

Yes multi-sig would work for this and it is supported by the protocol (although incomplete at this point).  However more likely companies would employ similar technology that they do now to protect PKI private keys.  The use of hardware security modules could allow business rules to be enforced at an enterprise level while not allowing any employee to ever have access to the actual private key.  This would decouple access from the actual private key.  Removing employee access could done via normal PKI processes (for example sending access revocation command to the HSM).

Also see, Sahai-Waters attribute based encryption:

http://eprint.iacr.org/2008/290.pdf