Wallet protection ideas for the super paranoid...

[mewantsbitcoins]

Who me? I don't even use truecrypt. All my drives got formatted long time and that thing that looks like random data - it's just elecromagnetic field from my crt playing tricks on carelessly placed HDD's. Hell, I wouldn't even be here if my neighbor wasn't so kind to lend me his laptop once in a while

[m4rkiz]

SSD wear leveling and the way free space must be prepared is an extra level of protection.  See this article about protection:
http://news.techworld.com/security/3263093/ssd-fimware-destroys-digital-evidence-researchers-find/

garbage collection - trim - only works when specific criteria are met, you need to be sure that it is actually working and it simply may not clear data you want to protect, while
one hdd overwrite cycle is enough

USB sticks are rated for around 10 or so years.

MLC chips, not whole pendrive!

In an air tight PVC pipe underground, a memory stick is in a better environment than above ground.  Above ground, you have more extreme temperature and humidity fluctuation.

check what happens with tin at low temperatures - α-tin

Granted.  My approach lacks redundancy

say you just used 20 new pendrives... from batch that was recalled later because of technical problems... you must have diversity there

With your method, you have just exposed your BitCoins to seizure by the state.  With my method, there would be nothing encrypted on my computer.

this is fairly small amount of data and can be hidden in number of ways, tc containers are not distinguishable from random data

[gigabytecoin]

If you hear voices often, this method is for you! 

[Enky1974]

You can put your encrypted wallet.dat inside a jpeg file, tt’s a pretty cool trick

[Zerbie]

If you hear voices often, this method is for you! 

How did you know?

[mewantsbitcoins]

If you hear voices often, this method is for you! 

The truly scary part to that is when you realize they are not only real, but also accurate

[Timo Y]

The problem with truecrypt is that everyone knows about those hidden volumes and courts might not be so easily convinced that there is no such hidden volume. This makes truecrypt style plausible deniability not so plausible in comparison with proposed above "decoy swarm" method.

But yes, truecrypt hidden volume is still something to consider, maybe even in combination with "decoy swarm" method.

Solution: Nested hidden volumes. The question is: How far must you nest until courts believe they have really arrived at the last layer?

I'd say 100 times at least 

[Timo Y]

Suggestions, comments?

1. Extract private key from wallet.dat

2. Split data into 100 equal chunks.

3. Encode each chunk in a human-memory friendly format.  You know, like 01101000101011 translating to "Pink elephant telephones dirty saxophone" or something like that.

4. Ask 300 of your friends to memorize one chunk each.  Each chunk is memorized by 3 friends for the sake of redundancy.  They are distributed in an order only know to you to ensure they will be reassembled correctly.  Perhaps you could offer your friends a small monetary award for correctly reciting their chunk once a year. Correctness is verified by comparing with the 2 other friends who memorized the same chunk.

5. Memorize the order in which your friends were assigned the chunks and destroy all physical traces of it.

6. Destroy all traces of wallet.dat and private key

[byronbb]

How about running the client on a VPS? The hosting service a) backs up regularly b) is more secure than your home computer c) is anonymous since no attacker would know to look there, and d) can be secured with iptables and passwords.

[fergalish]

Even better idea is to use truecrypt
http://www.truecrypt.org/docs/?s=plausible-deniability

Once upon a time there was also rubberhose and phonebookfs.  They were well designed for plausible deniability.  See http://en.wikipedia.org/wiki/Deniable_encryption#Software  Sadly never reached release state and long since discontinued.