🚀 go-pugleaf

"Clemens" <metax@despammed.com> writes:

> dd if=urandom of=/dev/hdb
> dd if=random of=/dev/hdb
> 
> I want to overwrite my harddisk with random gibberish.
> What are the effects if I use "urandom"?
> What are the effects if I use "random"?
> 
> What would you adwise, what�s more reliable and secure?
> 

Here is the extract from urandom(4) man page on my Debian Linux:

       When read, the /dev/random device will only return  random
       bytes  within the estimated number of bits of noise in the
       entropy pool.  /dev/random should  be  suitable  for  uses
       that  need  very  high quality randomness such as one-time
       pad or key generation.  When the entropy  pool  is  empty,
       reads  to /dev/random will block until additional environ�
       mental noise is gathered.

       When read, /dev/urandom device will return as  many  bytes
       as are requested.  As a result, if there is not sufficient
       entropy in the entropy pool, the returned values are theo�
       retically  vulnerable  to  a  cryptographic  attack on the
       algorithms used by the driver.  Knowledge  of  how  to  do
       this is not available in the current non-classified liter�
       ature, but it  is  theoretically  possible  that  such  an
       attack  may  exist.  If this is a concern in your applica�
       tion, use /dev/random instead.


In other words, /dev/random is more secure but much slower and will most
certainly block when large number of data is required (I think somebody
posted that he needed 10 minutes with /dev/random for the same amount of
data that  /dev/urandom could generated within 10sec).

Now, both of these might be an overkill for you.

HTH, Dragan

-- 
Dragan Cvetkovic, 

To be or not to be is true. G. Boole      No it isn't.  L. E. J. Brouwer

In article <3d7cd0f2$0$25296$91cee783@newsreader01.highway.telekom.at>,
"Clemens" <metax@despammed.com> wrote:

> dd if=urandom of=/dev/hdb
> dd if=random of=/dev/hdb
> 
> I want to overwrite my harddisk with random gibberish. What are the
> effects if I use "urandom"? What are the effects if I use "random"?
> 
> What would you adwise, what�s more reliable and secure?
> 
> 

Why not use both? I assume your doing this to "shred" your data. If so,
do both like 5 times over, and maybe the FBI won't be able to find your
info. Do it like 2 million times, and you'll be sure that the FBI isn't
after you, cause it'll probably take you like 2million days to shred the
/dev/hdb that much 0.o

PS: always set a followup-to header when posting to more then one group

Centuries ago, Nostradamus foresaw when "Clemens" <metax@despammed.com> would write:
> dd if=urandom of=/dev/hdb
> dd if=random of=/dev/hdb
>
> I want to overwrite my harddisk with random gibberish.
> What are the effects if I use "urandom"?
> What are the effects if I use "random"?
>
> What would you adwise, what�s more reliable and secure?

/dev/random will block any time the "entropy pool" runs out of bits
that it considers random, and wait until it gets some more entropy.
It will take a HUGE amount of time to run on a sizable partition.

/dev/urandom does not block in that manner, and will simply use the
cipher feedback to rehash whatever is in the entropy pool.

I would suggest a _third_ possibility, namely that you set up your own
cryptographically-strong RNG (there's lots of them out there to choose
from), using it to generate data to use as "gibberish," periodically
grabbing a byte or two from /dev/urandom or /dev/random to make it
behave a bit more randomly.

In addition, I'd suggest running this process multiple times, as it is
reputed that Those TLA Agencies are still able to pull data off if you
have only written over the disk once.

Of course, if you're really _serious_ about making this secure, you'd
want to use thermite to burn up the disk, grind it to powder, and then
mix it with thermite again...
-- 
(concatenate 'string "cbbrowne" "@ntlug.org")
http://cbbrowne.com/info/
Time is the best teacher; Unfortunately it kills all its students.

Clemens wrote:
> 
> dd if=urandom of=/dev/hdb
> dd if=random of=/dev/hdb
> 
> I want to overwrite my harddisk with random gibberish.
> What are the effects if I use "urandom"?
> What are the effects if I use "random"?
> 
> What would you adwise, what�s more reliable and secure?

Using /dev/random for this is completely infeasible. So I
advice you to use urandom if you want to use one of them.

-- 
Kasper Dupont -- der bruger for meget tid p� usenet.
For sending spam use mailto:aaarep@daimi.au.dk
or mailto:mcxumhvenwblvtl@skrammel.yaboo.dk

Dragan Cvetkovic wrote:
> 
> In other words, /dev/random is more secure but much slower and will most
> certainly block when large number of data is required (I think somebody
> posted that he needed 10 minutes with /dev/random for the same amount of
> data that  /dev/urandom could generated within 10sec).

We discussed that back in february: <news:3C6EE1AA.6ECB954E@daimi.au.dk>

The speed of /dev/random is mostly affected by the speed of the user.
While the speed of /dev/urandom is only affected by the speed of the
CPU. So comparing them will give different results depnding on the
speed of your CPU and the speed of your user.

If I move my mouse like if I was mad, I am able to produce 10KB/s
from /dev/random. OTOH /dev/urandom produces 440KB/s with my 350MHz
AMD K6/2.

Others have expirenced larger differences like 1KB/s vs. 1MB/s. The
speed of randomness from the mouse might depend on the type, at
least it seemed to be the case that PS/2 mice gives more randomness
than serial mice.

To wipe a 10GB harddisk with /dev/random, you would have to move the
mouse for 11 days nonstop. And you should repeat that three times.
That is more than one month of nonstop mousemovements. I think if
you do that you will have something else to worry about than if
anybody is able to read your disk.

-- 
Kasper Dupont -- der bruger for meget tid p� usenet.
For sending spam use mailto:aaarep@daimi.au.dk
or mailto:mcxumhvenwblvtl@skrammel.yaboo.dk

Clemens wrote:

> dd if=urandom of=/dev/hdb
> dd if=random of=/dev/hdb
> 
> I want to overwrite my harddisk with random gibberish.

Why not simply mount it as /var/spool/news for a few days?  *wink*
 
> What would you adwise, what�s more reliable and secure?

Actually for your application, wiping a hard disk absolutely clean I 
assume, it'll be the number of passes that prevent specialist analysis of 
what was on the disk previously.

I would use 'badblocks -p 4 -w /dev/hdb' which makes sure each bit on the 
drive is flipped multiple times, and then overwrite with /dev/urandom to 
fill it with gibberish.

You should then zero out the first block to create a correct absent 
partition table.

Rob

>>>>> "Christopher" == Christopher Browne <cbbrowne@acm.org> writes:

    Christopher> I would suggest a _third_ possibility, namely that
    Christopher> you set up your own cryptographically-strong RNG
    Christopher> (there's lots of them out there to choose from),
    Christopher> using it to generate data to use as "gibberish,"
    Christopher> periodically grabbing a byte or two from /dev/urandom
    Christopher> or /dev/random to make it behave a bit more randomly.

For seeding an RNG and doing it just occasionally, I think /dev/random
would be better.   (Are there any good reasons  to use /dev/urandom in
this case?)


    Christopher> In addition, I'd suggest running this process
    Christopher> multiple times, as it is reputed that Those TLA
    Christopher> Agencies are still able to pull data off if you have
    Christopher> only written over the disk once.

Yeah!  That's  why I  think using /dev/urandom  for 6 passes  would be
better than doing /dev/random 3 times.  And the former is much faster!



-- 
Lee Sau Dan                     ���u��(Big5)                    ~{@nJX6X~}(HZ) 

E-mail: danlee@informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee

Hi,

I do not think that the random/encryption algo make any difference. The
attacker is interested in what is not visible. The attacker will try to
subtrack the overwriting data form the signal, and in this case it does not
matter what the overwriting signal was, because it is clearly readable.

The signal were they are after, is the magnetic field that is slightly
outside the track, caused by head placement tolerance or diffusion.

The old algorithmn of writing all 1's thane all 0's and than a fixed pattern
is ment to counter the effect of magnetic hysterese.

Complete erasure is difficult, I have seen a report were they had to use AC
fields
followed by DC fields from a superconducting magnet. And that was before GMR
heads

Wim

Wim Ton wrote:
> 
> Hi,
> 
> I do not think that the random/encryption algo make any difference. The
> attacker is interested in what is not visible. The attacker will try to
> subtrack the overwriting data form the signal, and in this case it does not
> matter what the overwriting signal was, because it is clearly readable.
> 
> The signal were they are after, is the magnetic field that is slightly
> outside the track, caused by head placement tolerance or diffusion.
> 
> The old algorithmn of writing all 1's thane all 0's and than a fixed pattern
> is ment to counter the effect of magnetic hysterese.
> 
> Complete erasure is difficult, I have seen a report were they had to use AC
> fields
> followed by DC fields from a superconducting magnet. And that was before GMR
> heads
> 
> Wim

I always liked the "drop it from the roof" technique myself  :)

-- 
David Filion		filiond[at]videotron[dot]ca
Linux counter #195143

"Yes, it's manly, but let's face it, so is bungee-jumping with the
 cord tied to your testicles." -- Linus Torvalds