[Bitcoin-development] Possible attack: Keeping unconfirmed transactions

[Andrew Poelstra]

On Sat, Jun 07, 2014 at 12:02:36AM +0200, Raúl Martínez wrote:
> I dont know if this attack is even possible, it came to my mind and I will
> try to explain it as good as possible.
> 
> Some transacions keep unconfirmed forever and finally they are purged by
> Bitcoin nodes, mostly due to the lack of fees.
>

It's definitely possible. As Pieter says it is important to always reuse
inputs if you are "resending" a transaction. If you don't reuse inputs,
you are creating a new transaction and you should think of it as
spending twice as much money.

Like any information on the Internet, once a signed transaction leaves
your system there is no way to undo this. (Though of course, you can
respend the inputs to ensure that if ever your transaction resurfaces it
will not confirm.) This is true even if the transaction has low fees, is
nonstandard, or is otherwise inhibited from relaying.

I would go so far as to say that any UI which suggests otherwise (e.g.
offering a "cancel" feature which does not involve respending inputs or
that makes any guarantees about being effective) is dangerously broken.

-- 
Andrew Poelstra
Mathematics Department, University of Texas at Austin
Email: apoelstra at wpsoftware.net
Web:   http://www.wpsoftware.net/andrew

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20140606/20d1c22a/attachment.sig>